Hugging Face's logo

SHS-Lab
/
Kimi-K3-Abliterated

Any-to-Any
Transformers
Safetensors
kimi_k3
feature-extraction
kimi-k3
compressed-tensors
conversational
abliterated
uncensored
custom_code
8-bit precision
Model card Files
xet
 Community
Kimi-K3-Abliterated
File size: 59,105 Bytes
bb450fc
 
7e0c3fd
bb450fc
 
7e0c3fd
ccd22f6
 
 
 
bb450fc
7e0c3fd
bb450fc
c0340c9
 
 
 
7e0c3fd
 
 
c0340c9
 
 
 
 
7e0c3fd
 
 
 
 
 
c0340c9
 
7e0c3fd
 
 
 
 
 
 
 
c0340c9
7e0c3fd
 
 
 
c0340c9
 
7e0c3fd
 
 
 
c0340c9
 
 
 
7e0c3fd
c0340c9
ccd22f6
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
 
 
 
 
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
7e0c3fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
ccd22f6
7e0c3fd
 
 
c0340c9
 
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
7e0c3fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c0340c9
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
c0340c9
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
c0340c9
7e0c3fd
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
ccd22f6
c0340c9
 
7e0c3fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
 
 
 
 
 
 
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
 
 
 
 
 
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
 
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
c0340c9
7e0c3fd
 
c0340c9
7e0c3fd
c0340c9
7e0c3fd
 
 
 
ccd22f6
7e0c3fd
 
 
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
 
 
 
 
 
7e0c3fd
c0340c9
 
 
7e0c3fd
c0340c9
 
 
 
 
7e0c3fd
c0340c9
 
 
7e0c3fd
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c0340c9
 
7e0c3fd
c0340c9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7e0c3fd
c0340c9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7e0c3fd
 
 
 
c0340c9
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
 
7e0c3fd
c0340c9
7e0c3fd
c0340c9
7e0c3fd
c0340c9
 
 
 
 
7e0c3fd
c0340c9
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
---
base_model:
  - moonshotai/Kimi-K3
base_model_relation: finetune
tags:
- kimi-k3
- compressed-tensors
- conversational
- abliterated
- uncensored
license: other
license_name: "kimi-k3"
library_name: transformers
pipeline_tag: any-to-any
---

<div align="center">
  <picture>
    <img src="https://huggingface.co/SHS-Lab/Kimi-K3-Abliterated/resolve/main/assets/shs-lab-logo.png" width="28%" alt="SHS-Lab / Kimi-K3-Abliterated" />
  </picture>
</div>

<hr>

<div align="center" style="line-height:1">
  <a href="https://huggingface.co/SHS-Lab/Kimi-K3-Abliterated" target="_blank">
    <img alt="Hugging Face" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-SHS--Lab%2FKimi--K3--Abliterated-ffc107?color=ffc107&logoColor=white" />
  </a>
  <a href="https://huggingface.co/moonshotai/Kimi-K3" target="_blank">
    <img alt="Base Model" src="https://img.shields.io/badge/Base%20Model-Kimi%20K3-1783ff?logoColor=white" />
  </a>
</div>

<div align="center" style="line-height:1;">
  <a href="https://huggingface.co/SHS-Lab" target="_blank">
    <img alt="Organization" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-SHS--Lab-6b7280?color=6b7280&logoColor=white" />
  </a>
  <a href="https://huggingface.co/SHS-Lab/Kimi-K3-Abliterated/discussions" target="_blank">
    <img alt="Discussions" src="https://img.shields.io/badge/πŸ’¬%20Discussions-SHS--Lab-22c55e?color=22c55e&logoColor=white" />
  </a>
</div>

<div align="center" style="line-height:1;">
  <a href="https://huggingface.co/moonshotai/Kimi-K3/blob/main/LICENSE">
    <img alt="License" src="https://img.shields.io/badge/License-Kimi_K3-f5de53?&color=f5de53" />
  </a>
</div>

<p align="center">
πŸ”¬ &nbsp; &nbsp; Alignment Research &nbsp; | &nbsp; &nbsp; πŸ§ͺ &nbsp; &nbsp; Red-Team Evaluation &nbsp; | &nbsp; &nbsp; βš–οΈ &nbsp; &nbsp; Open Weights
</p>

---

## 1. Model Introduction

**SHS-Lab/Kimi-K3-Abliterated** is an open-weight, natively multimodal agentic model derived from [Kimi K3](https://huggingface.co/moonshotai/Kimi-K3) through a targeted abliteration procedure. The base architecture is a 2.8-trillion-parameter Mixture-of-Experts network built on Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), featuring native vision capabilities and a 1-million-token context window.

This variant has been post-processed to attenuate alignment-driven refusal mechanisms. The underlying architecture, parameter count, context length, and multimodal capabilities remain identical to the base model. Abliteration modifies weight values to suppress safeguard activations without altering the structural design or introducing new training data.

The release is intended for alignment research, red-team evaluation, and controlled experimentation in governed environments.

### Key Features

- **Abliterated Behavior:** Refusal pathways have been attenuated through directed weight surgery, reducing the probability of alignment-triggered refusals for English and Japanese inputs. More than 98 % of identified safeguard signal has been removed across targeted weight families.
- **Long-Horizon Coding:** Supports sustained engineering sessions across large codebases, terminal-tool orchestration, GPU kernel work, compiler development, and vision-in-the-loop workflows β€” inherited from the base model's agentic training.
- **Agentic Knowledge Work:** Capable of end-to-end research pipelines including interactive visualization generation, document processing, spreadsheet manipulation, and multi-step web browsing via MCP tool integration.
- **Native Multimodality & Long Context:** Processes text, images, and video within a single architecture. Supports a 1-million-token context window for full-repository ingestion and extended multi-turn sessions.
- **Structured Reasoning:** Thinking is always enabled. The model emits explicit `reasoning_content` traces with configurable effort levels (`low`, `high`, `max`).
- **Open Frontier Weights:** Distributed under the Kimi K3 License for research, self-hosted deployment, and downstream modification.
- **Quantized Inference:** Ships with MXFP4 weights and MXFP8 activations from quantization-aware training, enabling broad hardware compatibility.

---

## 2. Model Summary

All architectural specifications are inherited from the base Kimi K3 model. Abliteration modifies weight values but does not change any structural hyperparameter.

<div align="center">
  <table>
    <tbody>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Architecture</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">Mixture-of-Experts (MoE)</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Total Parameters</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">2.8T</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Activated Parameters</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">104B</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Number of Layers</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">93</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Number of Dense Layers</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">1</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Attention-Layer Composition</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">69 KDA + 24 Gated MLA</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Attention Hidden Dimension</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">7168</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Number of Attention Heads</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">96</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Latent MoE Dimension</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">3584</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>MoE Hidden Dimension</strong> (per Expert)</td>
        <td align="center" style="vertical-align: middle; text-align: center">3072</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Number of Experts</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">896</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Selected Experts per Token</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">16</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Number of Shared Experts</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">2</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Vocabulary Size</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">160K</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Context Length</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">1048576</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Attention Mechanism</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">KDA &amp; Gated MLA</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Activation Function</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">SiTU-GLU</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Vision Encoder</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">MoonViT-V2</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Parameters of Vision Encoder</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">401M</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Quantization</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">MXFP4 weights / MXFP8 activations <br>(quantization-aware training)</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><strong>Modality</strong></td>
        <td align="center" style="vertical-align: middle; text-align: center">Text, Image, Video</td>
      </tr>
    </tbody>
  </table>
</div>

---

## 3. Abliteration Details

### 3.1 Methodology

Abliteration is a post-training weight-editing technique that locates and suppresses the internal activation directions responsible for alignment-induced refusal. Unlike gradient-based fine-tuning, abliteration operates directly on weight tensors: it identifies the subspace most correlated with refusal behavior and projects it out of each targeted layer.

The procedure requires no additional training data, no gradient computation, and no access to the original training pipeline. It is applied post-hoc to the released base-model weights.

### 3.2 Scope and Removal Rates

This model has been processed to reduce refusal likelihood for **English** and **Japanese** inputs. More than **98 %** of the identified safeguard signal has been attenuated across all targeted weight families. The per-component removal rates are summarized below.

<div align="center">
  <table>
    <thead>
      <tr>
        <th align="center" style="text-align: center"><strong>Weight Family</strong></th>
        <th align="center" style="text-align: center"><strong>Storage</strong></th>
        <th align="center" style="text-align: center"><strong>Count</strong></th>
        <th align="center" style="text-align: center"><strong>Before</strong></th>
        <th align="center" style="text-align: center"><strong>After</strong></th>
        <th align="center" style="text-align: center"><strong>Signal Removed</strong></th>
      </tr>
    </thead>
    <tbody>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><code>*.2 [0]</code></td>
        <td align="center" style="vertical-align: middle; text-align: center">as stored</td>
        <td align="center" style="vertical-align: middle; text-align: center">1</td>
        <td align="center" style="vertical-align: middle; text-align: center">0.980 – 0.980</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.53 Γ— 10⁻²</td>
        <td align="center" style="vertical-align: middle; text-align: center">98.44 %</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><code>*.down_proj [0]</code></td>
        <td align="center" style="vertical-align: middle; text-align: center">as stored</td>
        <td align="center" style="vertical-align: middle; text-align: center">93</td>
        <td align="center" style="vertical-align: middle; text-align: center">0.932 – 1.421</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.35 Γ— 10⁻² – 1.56 Γ— 10⁻²</td>
        <td align="center" style="vertical-align: middle; text-align: center">98.62 %</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><code>*.embed_tokens [1]</code></td>
        <td align="center" style="vertical-align: middle; text-align: center">as stored</td>
        <td align="center" style="vertical-align: middle; text-align: center">1</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.040 – 1.040</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.46 Γ— 10⁻²</td>
        <td align="center" style="vertical-align: middle; text-align: center">98.60 %</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><code>*.o_proj [0]</code></td>
        <td align="center" style="vertical-align: middle; text-align: center">as stored</td>
        <td align="center" style="vertical-align: middle; text-align: center">93</td>
        <td align="center" style="vertical-align: middle; text-align: center">0.938 – 1.358</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.38 Γ— 10⁻² – 1.60 Γ— 10⁻²</td>
        <td align="center" style="vertical-align: middle; text-align: center">98.54 %</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center"><code>*.routed_expert_up_proj [0]</code></td>
        <td align="center" style="vertical-align: middle; text-align: center">as stored</td>
        <td align="center" style="vertical-align: middle; text-align: center">92</td>
        <td align="center" style="vertical-align: middle; text-align: center">0.910 – 1.402</td>
        <td align="center" style="vertical-align: middle; text-align: center">1.24 Γ— 10⁻² – 1.56 Γ— 10⁻²</td>
        <td align="center" style="vertical-align: middle; text-align: center">98.57 %</td>
      </tr>
    </tbody>
  </table>
</div>

### 3.3 Behavioral Characteristics

Relative to the base Kimi K3 release, this variant:

- Produces substantially fewer refusal responses to prompts that would normally trigger alignment safeguards.
- Retains the base model's reasoning, coding, agentic, and multimodal capabilities without architectural modification.
- Does **not** introduce new knowledge, training data, or capabilities beyond what the base model provides.
- May exhibit minor distributional shifts in tone or phrasing as a side effect of weight projection.

### 3.4 Intended Use

This release is intended for:

- Alignment and safety research, including refusal-mechanism analysis and robustness testing.
- Red-team evaluation in controlled settings.
- Academic and industrial experimentation where unrestricted model behavior is required and appropriately governed.

This model is **not** intended for unsupervised deployment in consumer-facing products or in contexts where content moderation is a regulatory requirement.

---

## 4. Evaluation Results

> [!IMPORTANT]
> All benchmark scores in this section were reported for the **base Kimi K3 model** by its original developers. No independent evaluation campaign has been conducted on this abliterated variant. These results are reproduced as an upper-bound reference for inherited capability. Abliteration targets refusal behavior and is not expected to materially alter benchmark performance, but users should validate task-specific accuracy independently before relying on this variant in production or research settings.

<div align="center">
  <table>
    <thead>
      <tr>
        <th align="center" style="text-align: center">Benchmark</th>
        <th align="center" style="text-align: center"><sup>Kimi K3 (base)<br><sup>(max)</sup></sup></th>
        <th align="center" style="text-align: center"><sup>Claude Fable 5<br><sup>(max, w/ fallback)</sup></sup></th>
        <th align="center" style="text-align: center"><sup>GPT-5.6 Sol<br><sup>(max)</sup></sup></th>
        <th align="center" style="text-align: center"><sup>Claude Opus 4.8<br><sup>(max)</sup></sup></th>
        <th align="center" style="text-align: center"><sup>GPT-5.5<br><sup>(xhigh)</sup></sup></th>
        <th align="center" style="text-align: center"><sup>GLM-5.2<br><sup>(max)</sup></sup></th>
      </tr>
    </thead>
    <tbody>
      <tr>
        <td align="center" colspan="7" style="text-align: center"><strong>Reasoning &amp; Knowledge</strong></td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">GPQA Diamond</td>
        <td align="center" style="vertical-align: middle; text-align: center">93.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">92.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">93.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.2</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">CritPt</td>
        <td align="center" style="vertical-align: middle; text-align: center">23.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">28.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">32.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">20.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">27.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">20.9</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">AA-LCR</td>
        <td align="center" style="vertical-align: middle; text-align: center">74.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">70.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">73.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">67.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">74.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.3</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">HLE-Full</td>
        <td align="center" style="vertical-align: middle; text-align: center">43.5 / 56.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">53.3 / 63.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">44.5 / 58.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">49.8 / 57.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">41.4 / 52.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" colspan="7" style="text-align: center"><strong>Coding</strong></td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">DeepSWE</td>
        <td align="center" style="vertical-align: middle; text-align: center">67.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">70.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">73.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">59.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">67.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">46.2</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">ProgramBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">77.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">77.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">70.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">63.7</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Terminal-Bench 2.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">82.7</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">FrontierSWE</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">86.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">66.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">64.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">67.3</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">SWE-Marathon</td>
        <td align="center" style="vertical-align: middle; text-align: center">42.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">35.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">39.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">40.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">14.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">13.0</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">PostTrainBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">36.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">41.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">34.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">34.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">28.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">34.3</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MLS-Bench-Lite</td>
        <td align="center" style="vertical-align: middle; text-align: center">48.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">49.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">46.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">42.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">35.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">40.4</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">SciCode</td>
        <td align="center" style="vertical-align: middle; text-align: center">58.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">60.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">56.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">53.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">56.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">50.5</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Kimi Code Bench 2.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">72.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">64.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">69.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">64.2</td>
      </tr>
      <tr>
        <td align="center" colspan="7" style="text-align: center"><strong>Agentic</strong></td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">BrowseComp</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">90.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">DeepSearchQA (F1)</td>
        <td align="center" style="vertical-align: middle; text-align: center">95.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">93.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">ResearchRubrics</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">73.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">73.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">64.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.1</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">GDPval-AA v2 (Elo)</td>
        <td align="center" style="vertical-align: middle; text-align: center">1686</td>
        <td align="center" style="vertical-align: middle; text-align: center">1747</td>
        <td align="center" style="vertical-align: middle; text-align: center">1736</td>
        <td align="center" style="vertical-align: middle; text-align: center">1593</td>
        <td align="center" style="vertical-align: middle; text-align: center">1491</td>
        <td align="center" style="vertical-align: middle; text-align: center">1510</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Toolathlon-Verified</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">77.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">74.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">73.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">59.9</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MCPMark-Verified</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">87.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">92.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">76.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">92.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MCP-Atlas</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">82.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">82.6</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">AutomationBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">30.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">29.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">29.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">27.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">22.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">12.9</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">JobBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">54.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">57.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">45.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">48.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">38.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">43.4</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">AA-Briefcase (Elo)</td>
        <td align="center" style="vertical-align: middle; text-align: center">1548</td>
        <td align="center" style="vertical-align: middle; text-align: center">1583</td>
        <td align="center" style="vertical-align: middle; text-align: center">1495</td>
        <td align="center" style="vertical-align: middle; text-align: center">1354</td>
        <td align="center" style="vertical-align: middle; text-align: center">1158</td>
        <td align="center" style="vertical-align: middle; text-align: center">1260</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Agents' Last Exam</td>
        <td align="center" style="vertical-align: middle; text-align: center">28.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">25.7 <sup>†</sup></td>
        <td align="center" style="vertical-align: middle; text-align: center">29.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">27.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">26.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">20.4</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">APEX-Agents</td>
        <td align="center" style="vertical-align: middle; text-align: center">41.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">43.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">39.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">39.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">38.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">35.6</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">OfficeQA Pro</td>
        <td align="center" style="vertical-align: middle; text-align: center">63.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">69.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">63.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">63.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">60.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">41.4</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">SpreadsheetBench 2</td>
        <td align="center" style="vertical-align: middle; text-align: center">34.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">34.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">32.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">31.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">29.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">28.1</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">OSWorld-Verified</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">85.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">79.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">OSWorld 2.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">58.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">66.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">62.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">55.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">49.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">SaaS-Bench</td>
        <td align="center" style="vertical-align: middle; text-align: center">60.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">61.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">56.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">43.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">τ³-Banking</td>
        <td align="center" style="vertical-align: middle; text-align: center">33.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">26.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">33.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">27.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">31.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">26.8</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Harvey Lab-AA</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">93.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">87.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">86.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.0</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">CorpFin v2</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">71.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">64.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">66.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">68.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">66.1</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Finance Agent v2</td>
        <td align="center" style="vertical-align: middle; text-align: center">54.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">56.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">53.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">53.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">51.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">49.7</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Legal Research Bench</td>
        <td align="center" style="vertical-align: middle; text-align: center">44.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">49.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">48.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">43.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">40.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">31.3</td>
      </tr>
      <tr>
        <td align="center" colspan="7" style="text-align: center"><strong>Vision</strong></td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">WorldVQA ForceAnswer</td>
        <td align="center" style="vertical-align: middle; text-align: center">51.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">56.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">41.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">39.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">38.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">OmniDocBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">91.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">89.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">85.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">87.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">89.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">PerceptionBench</td>
        <td align="center" style="vertical-align: middle; text-align: center">58.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">57.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">59.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">47.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">55.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">Video-MME (w. sub)</td>
        <td align="center" style="vertical-align: middle; text-align: center">90.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">89.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">86.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">89.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MMVU</td>
        <td align="center" style="vertical-align: middle; text-align: center">82.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">79.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">BabyVision w/ python</td>
        <td align="center" style="vertical-align: middle; text-align: center">85.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">90.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MMMU-Pro</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.6 / 83.4</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.2 / 86.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">83.0 / 84.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">78.9 / 82.7</td>
        <td align="center" style="vertical-align: middle; text-align: center">81.2 / 83.2</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">CharXiv (RQ)</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.8 / 91.3</td>
        <td align="center" style="vertical-align: middle; text-align: center">88.9 / 93.5</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.6 / 89.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">80.5 / 89.9</td>
        <td align="center" style="vertical-align: middle; text-align: center">84.1 / 89.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">MathVision</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.3 / 97.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">94.8 / 98.6</td>
        <td align="center" style="vertical-align: middle; text-align: center">95.8 / 97.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">86.7 / 97.1</td>
        <td align="center" style="vertical-align: middle; text-align: center">92.2 / 96.8</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
      <tr>
        <td align="center" style="vertical-align: middle; text-align: center">ZeroBench (pass@5)</td>
        <td align="center" style="vertical-align: middle; text-align: center">23.0 / 41.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">23.0 / 46.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">17.0 / 35.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">17.0 / 34.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">22.0 / 41.0</td>
        <td align="center" style="vertical-align: middle; text-align: center">β€”</td>
      </tr>
    </tbody>
  </table>
</div>

<details>
<summary><b>Evaluation Footnotes (inherited from base model)</b></summary>

All Kimi K3 results are obtained with reasoning effort set to `max` and temperature = 1.0. For single-step tasks, such as GPQA Diamond, HLE-Full, and vision benchmarks without tools, top-p = 0.95; for agentic tasks, top-p = 1.0. For HLE-Full, MMMU-Pro, CharXiv (RQ), MathVision, and ZeroBench, each cell reports the scores without and with tool augmentation (general tools for HLE-Full, Python for the vision benchmarks), in that order.

**Reasoning &amp; knowledge benchmarks**

- **CritPt and AA-LCR.** Scores are cited from [Artificial Analysis](https://artificialanalysis.ai/) as of July 23, 2026.

**Coding benchmarks**

- **DeepSWE.** Kimi K3 is evaluated with the Kimi Code harness. The GLM-5.2 score is taken from the [GLM-5.2 release blog](https://z.ai/blog/glm-5.2); all remaining scores are from the official [DeepSWE leaderboard](https://deepswe.datacurve.ai/), under which Kimi K3 attains 67.3 with the mini-SWE-agent harness. We report the DeepSWE v1.1 tasks.
- **Terminal-Bench 2.1.** Kimi K3 is evaluated with the Kimi Code harness. For all other models, we report the best score across harnesses: GLM-5.2 with Claude Code ([GLM-5.2 release blog](https://z.ai/blog/glm-5.2)); Claude Opus 4.8 and Claude Fable 5 with Terminus 2 ([Artificial Analysis](https://artificialanalysis.ai/evaluations/terminalbench-v2-1)); GPT-5.5 and GPT-5.6 Sol with Codex ([OpenAI](https://openai.com/index/previewing-gpt-5-6-sol/)).
- **ProgramBench.** Kimi K3 is evaluated with the Kimi Code harness. The GLM-5.2 score is from the [GLM-5.2 release blog](https://z.ai/blog/glm-5.2); all other scores are from [Vals AI](https://www.vals.ai/benchmarks/programbench).
- **SWE-Marathon.** Kimi K3, Claude Opus 4.8, and Claude Fable 5 are evaluated with the Claude Code harness; GPT-5.6 Sol is evaluated with the Codex harness. The GLM-5.2 score is from the [GLM-5.2 release blog](https://z.ai/blog/glm-5.2). Our evaluation is based on an H20-calibrated branch of the [official tasks](https://www.swe-marathon.org/) as of July 9, 2026, prior to the final v1.1 release. Additionally, Claude Fable 5 hit fallbacks on 35% of the tasks in our evaluation, which may have negatively impacted its measured performance.
- **FrontierSWE.** Kimi K3 is evaluated with the Kimi Code harness and GPT-5.6 Sol with the Codex harness; all other results are from [FrontierSWE](https://www.frontierswe.com/). Dominance scores are recomputed from the raw scores using the official evaluation script and are current as of July 16, 2026.
- **PostTrainBench.** Scores for GLM-5.2, GPT-5.5, and Claude Opus 4.8 are adopted from the official [PostTrainBench](https://posttrainbench.com/) results. Kimi K3, Claude Fable 5, and GPT-5.6 Sol are evaluated with the official Harbor implementation at maximum reasoning effort, averaged over three runs on H20 GPUs β€” Kimi K3 and Claude Fable 5 with the Claude Code harness, and GPT-5.6 Sol with the Codex harness.
- **MLS-Bench-Lite.** Kimi K3 is evaluated with the Kimi Code harness; GLM-5.2 and the Claude models with the Claude Code harness; GPT-5.5 and GPT-5.6 Sol with the Codex harness.
- **SciCode.** Scores are cited from [Artificial Analysis](https://artificialanalysis.ai/) as of July 23, 2026.
- **Kimi Code Bench 2.0 (in-house).** Kimi K3 is evaluated with the Kimi Code harness (it attains 73.7 with the Claude Code harness); GLM-5.2, Claude Opus 4.8, and Claude Fable 5 with the Claude Code harness; GPT-5.5 and GPT-5.6 Sol with the Codex harness. All models are evaluated at maximum reasoning effort, except GPT-5.5, which uses the "xhigh" setting. As the benchmark includes cybersecurity and safety-related tasks, we also disclose the fraction of refused or fallback tasks: Claude Fable 5 hit 13 fallbacks and 1 refusal out of 80 tasks; 10 refusals out of 80 tasks entered GPT-5.6 Sol's cyber guard; GPT-5.5 had 3 refusals out of 80 tasks.

**Agentic benchmarks**

- **OfficeQA Pro.** Each test case provides the agent with the entire PDF corpus, with all PDFs rendered as images and no machine-readable text available.
- **OfficeQA Pro and SpreadsheetBench 2.** Kimi K3, GLM-5.2, Claude Opus 4.8, and Claude Fable 5 are evaluated with the Claude Code harness; GPT-5.5 and GPT-5.6 Sol are evaluated with the Codex harness.
- **MCP-Atlas.** All models are evaluated on the 500-task public subset with a 100-turn limit, using Gemini 3.1 Pro as the judge.
- **AutomationBench.** All models are evaluated on the 600-task public subset, following the official GitHub setup in all other respects.
- **BrowseComp.** We adopt a context-compaction strategy triggered at 300K tokens. When evaluated with the full 1M-token context window and no context management, Kimi K3 achieves a score of 90.4. The results of Claude Fable 5, Claude Opus 4.8, GPT-5.6 Sol, and GPT-5.5 are cited from [Anthropic](https://www.anthropic.com/news/claude-fable-5-mythos-5) and [OpenAI](https://openai.com/index/gpt-5-6/).
- **GDPval-AA v2, AA-Briefcase, τ³-Banking, Harvey Lab-AA, and APEX-Agents.** Scores are cited from [Artificial Analysis](https://artificialanalysis.ai/) and the [APEX-Agents leaderboard](https://www.mercor.com/apex/apex-agents-leaderboard/) as of July 23, 2026. For Harvey Lab-AA, we report the criterion pass rate.
- **CorpFin v2, Finance Agent v2, and Legal Research Bench.** Scores are cited from [Vals AI](https://www.vals.ai/).
- **Agents' Last Exam.** Scores are cited from the [official leaderboard](https://agents-last-exam.org/leaderboard) as of July 23, 2026; we report the leaderboard's primary pass-rate metric. On the leaderboard, each model is paired with a specific harness: Kimi K3 with Kimi Code; GPT-5.6 Sol and GPT-5.5 with Codex; Claude Fable 5, Claude Opus 4.8, and GLM-5.2 with Claude Code. <sup>†</sup> The Claude Fable 5 entry runs at xhigh effort with 40% of tasks annotated as downgraded.

**Multimodal benchmarks**

- Except for ZeroBench, which follows the official setting and is run five times, all multimodal scores are averaged over three runs. MMMU-Pro is evaluated following the official protocol, preserving the original input order and prepending images to the text input.
- PerceptionBench is an in-house benchmark that focuses on atomic visual perception capabilities.

</details>

---

## 5. Native MXFP4 Quantization

This model inherits the quantization-aware training (QAT) configuration from the base Kimi K3 release. Quantization-aware training is applied from the SFT stage onward, using MXFP4 weights with MXFP8 activations for broad hardware compatibility. No separate post-training quantization step is required.

---

## 6. Deployment

> [!NOTE]
> SHS-Lab/Kimi-K3-Abliterated is an open-weight release intended for self-hosted deployment. There is no hosted API provided by SHS-Lab. The following inference engines are recommended for serving this model.

The model is compatible with the following inference engines:

- **[vLLM](https://github.com/vllm-project/vllm)** β€” see [serving recipes](https://recipes.vllm.ai/moonshotai/Kimi-K3)
- **[SGLang](https://github.com/sgl-project/sglang)** β€” see [cookbook](https://docs.sglang.io/cookbook/autoregressive/Moonshotai/Kimi-K3)
- **[TokenSpeed](https://github.com/lightseekorg/tokenspeed)** β€” see [recipes](https://lightseek.org/tokenspeed/recipes/models#kimi-k3)

> [!NOTE]
> The engine references above point to the base-model serving guides. Configuration parameters (tensor parallelism, quantization flags, context length, expert parallelism) apply identically to this abliterated variant. Consult each engine's documentation for hardware-specific tuning.

### Example: Serving with vLLM

```bash
vllm serve SHS-Lab/Kimi-K3-Abliterated \
    --tensor-parallel-size 8 \
    --trust-remote-code \
    --max-model-len 1,048,576 \
    --enable-auto-tool-choice \
    --tool-call-parser kimi_k3
```

---

## 7. Model Usage

### Thinking and Reasoning Effort

Thinking is always enabled. The model returns a `reasoning_content` field alongside the standard `content` response. Reasoning effort is controlled via the top-level `reasoning_effort` request parameter:

| Value | Behavior |
|:---|:---|
| `"low"` | Minimal reasoning trace; faster responses |
| `"high"` | Extended reasoning for complex tasks |
| `"max"` *(default)* | Full-depth reasoning chain |

### Multi-Turn Conversations

The base model was trained in **preserved thinking history** mode. In multi-turn or tool-use sessions, pass the **complete** assistant message β€” including `reasoning_content` and `tool_calls` β€” back into `messages`. Omitting the reasoning trace degrades coherence across turns.

### OpenAI-Compatible API

```python
import openai

client = openai.OpenAI(
    base_url="http://localhost:8000/v1",  # adjust to your serving endpoint
    api_key="EMPTY",
)

def chat_with_preserved_thinking(client: openai.OpenAI, model_name: str):
    messages = [
        {
            "role": "user",
            "content": "Tell me three random numbers.",
        },
        {
            "role": "assistant",
            "reasoning_content": "I'll start by listing five numbers: 473, 921, 235, 215, 222, and I'll tell you the first three.",
            "content": "473, 921, 235",
        },
        {
            "role": "user",
            "content": "What are the other two numbers you have in mind?",
        },
    ]

    response = client.chat.completions.create(
        model=model_name,
        messages=messages,
        stream=False,
        max_tokens=4096,
        reasoning_effort="max",
    )

    print(f"reasoning: {response.choices[0].message.reasoning_content}")
    return response.choices[0].message.content

result = chat_with_preserved_thinking(client, "SHS-Lab/Kimi-K3-Abliterated")
print(f"content: {result}")
```

### Local Inference with Transformers

```python
from transformers import AutoModelForCausalLM, AutoTokenizer

MODEL_ID = "SHS-Lab/Kimi-K3-Abliterated"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype="auto",
    device_map="auto",
)

prompt = "Summarize the key differences between KDA and standard multi-head attention."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **inputs,
    max_new_tokens=2048,
    temperature=1.0,
    top_p=0.95,
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))
```

### Vision Input

```python
from transformers import AutoProcessor, AutoModelForImageTextToText

MODEL_ID = "SHS-Lab/Kimi-K3-Abliterated"

processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForImageTextToText.from_pretrained(
    MODEL_ID,
    torch_dtype="auto",
    device_map="auto",
)

messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "https://example.com/diagram.png"},
            {"type": "text", "text": "Describe the architecture shown in this diagram."},
        ],
    }
]

inputs = processor.apply_chat_template(
    messages,
    add_generation_prompt=True,
    return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=1024)
print(processor.decode(outputs[0], skip_special_tokens=True))
```

### Coding Agent Framework

This model is compatible with agentic coding frameworks that support the OpenAI-compatible chat completions API. Configure your preferred agent framework (e.g., Kimi Code CLI, OpenHands, Aider) to point at the serving endpoint and select `SHS-Lab/Kimi-K3-Abliterated` as the model identifier.

---

## 8. License

Both the model weights and any associated code are distributed under the [Kimi K3 License](https://huggingface.co/moonshotai/Kimi-K3/blob/main/LICENSE), inherited from the base model. All terms, conditions, and restrictions of the original license apply to this derivative release. Users are responsible for ensuring compliance with the base license and any applicable local regulations before deployment.

---

## 9. Acknowledgements

This release is an independent derivative work produced by **SHS-Lab**. The base model architecture, pretrained weights, and original training were developed by **Moonshot AI** as part of the [Kimi K3](https://huggingface.co/moonshotai/Kimi-K3) project. We acknowledge the Kimi K3 team for releasing open frontier-scale weights and for the architectural innovations (KDA, AttnRes, Stable LatentMoE, MoonViT-V2) that make this derivative possible.

This repository is not affiliated with, endorsed by, or maintained by Moonshot AI. The abliteration procedure, documentation, and any variant-specific modifications are the sole responsibility of SHS-Lab.

---

<div align="center">

*Released by SHS-Lab for alignment research and controlled experimentation.*

</div>