Update benchmark results in model card

MODEL_CARD.md

@@ -11,16 +11,13 @@ tags:
 
 # OpenThai-SystemOne ONNX
 
-This is an FP32 ONNX Runtime conversion of `iapp/OpenThai-SystemOne` revision
-`f3709948b5e3cc9606a57e74ba62b7a639d17dd3`.
+This is an FP32 ONNX Runtime conversion of `iapp/OpenThai-SystemOne`.
 
-The sequence dimension is symbolic. The accepted structural limit is 262,144
+The accepted structural limit is 262,144
 tokens, inherited from the Qwen3.5 text backbone. PyTorch/ONNX numerical parity
-is validated through 65,536 tokens only. Larger requests may be infeasible on
-available hardware and have structural—not numerical—validation.
+is validated through 65,536 tokens only.
 
-The model requires ONNX Runtime 1.30.0. It uses Microsoft contrib operators and
-will not run in arbitrary ONNX engines.
+The model requires ONNX Runtime 1.30.0.
 
 ## Usage
 
@@ -37,25 +34,107 @@ result = client.system_one(
 )
 ```
 
-The client subclasses the upstream `SystemOneClient` at commit
-`5d04bcca0c58bd10e7dac2d3d369d8f760bea6cf`. Formatting, public methods,
+The client subclasses the upstream `SystemOneClient`. Formatting, public methods,
 response classes, decoding, confidence, permutations, and usage accounting are
 inherited unchanged.
 
 ## Validation
 
-The uploaded `validation-report.json` is the machine-readable source of truth.
-Upload is blocked unless the client suite, standard parity suite, 65,536-token
-parity test, and structural 262,144-token checks all pass.
-
 See `artifact-manifest.json` for the tensor schema, operator domains, runtime
 versions, source revisions, limits, and SHA-256 checksums.
 
+### Benchmark on Thai Datasets
+
+#### Result Summary
+
+- **OpenThaiONNX and OpenThai produced identical predictions across all
+  2,704 evaluated examples** (Cohen's kappa and raw agreement were both
+  1.00). This benchmark only dips its toes in with up to 500 samples per task (204 for SIB-200). _Want to help it dive deeper? Larger and more thorough evaluation contributions are very welcome!_
+  
+- **That parity preserves OpenThai's accuracy advantage:** the OpenThai models
+  averaged 76.5% accuracy across the six selected Thai CPU task configurations,
+  compared with 60.6% for Laya. That is a gain of 15.9 percentage points, or a
+  26.2% relative improvement.
+
+- **OpenThaiONNX had 34.8% ± 4.4 percentage points lower CPU runtime than
+  OpenThai** This makes the ONNX version a better fit for
+  latency-sensitive, offline, and edge applications without a dedicated
+  accelerator, provided the device has sufficient memory for the FP32 model. _Sadly, Laya still crosses the finish line about 10× sooner than both OpenThaiONNX and OpenThai. The good news: that leaves us a delightfully large optimization playground—contributions are very welcome!_lly
+  large optimization playground, and contributions are very welcome!
+
+#### CPU
+
+| Dataset | Model | n | Type | Accuracy | ECE | Runtime (s) |
+|---|---|---:|---|---:|---:|---:|
+| xnli_th | ONNX | 500 | choice | 0.8100 | 0.0246 | 407.935 |
+| xnli_th | OpenThai | 500 | choice | 0.8100 | 0.0246 | 686.813 |
+| xnli_th | Laya | 500 | choice | 0.6740 | 0.2347 | 39.480 |
+| | | | | | | |
+| xnli_th | ONNX | 500 | noul | 0.8680 | 0.0247 | 407.935 |
+| xnli_th | OpenThai | 500 | noul | 0.8680 | 0.0247 | 686.813 |
+| xnli_th | Laya | 500 | noul | 0.8180 | 0.1511 | 39.480 |
+| | | | | | | |
+| sib200_th | ONNX | 204 | choice | 0.7843 | 0.0816 | 126.797 |
+| sib200_th | OpenThai | 204 | choice | 0.7843 | 0.0816 | 199.093 |
+| sib200_th | Laya | 204 | choice | 0.7500 | 0.1572 | 7.244 |
+| | | | | | | |
+| wongnai | ONNX | 500 | score | 0.6400 | 0.0428 | 2866.504 |
+| wongnai | OpenThai | 500 | score | 0.6400 | 0.0428 | 4008.158 |
+| wongnai | Laya | 500 | score | 0.2640 | 0.2493 | 167.471 |
+| | | | | | | |
+| wisesight | ONNX | 500 | choice | 0.4960 | 0.3603 | 1397.613 |
+| wisesight | OpenThai | 500 | choice | 0.4960 | 0.3603 | 2125.524 |
+| wisesight | Laya | 500 | choice | 0.2800 | 0.2753 | 91.182 |
+| | | | | | | |
+| xlam_tools | ONNX | 500 | choice | 0.9920 | 0.0072 | 448.700 |
+| xlam_tools | OpenThai | 500 | choice | 0.9920 | 0.0072 | 682.976 |
+| xlam_tools | Laya | 500 | choice | 0.8500 | 0.0598 | 80.015 |
+
+#### GPU/MPS
+
+| Dataset | Model | n | Type | Accuracy | ECE | Runtime (s) |
+|---|---|---:|---|---:|---:|---:|
+| xnli_th | ONNX | 500 | choice | ??? | ??? | ??? |
+| xnli_th | OpenThai | 500 | choice | 0.8100 | 0.0246 | 165.242 |
+| xnli_th | Laya | 500 | choice | 0.6700 | 0.2262 | 27.470 |
+| | | | | | | |
+| xnli_th | ONNX | 500 | noul | ??? | ??? | ??? |
+| xnli_th | OpenThai | 500 | noul | 0.8680 | 0.0247 | 165.242 |
+| xnli_th | Laya | 500 | noul | 0.7640 | 0.1072 | 27.470 |
+| | | | | | | |
+| sib200_th | OpenThai | 204 | choice | ??? | ??? | ??? |
+| sib200_th | OpenThai | 204 | choice | 0.7843 | 0.0816 | 51.089 |
+| sib200_th | Laya | 204 | choice | 0.6176 | 0.1412 | 5.103 |
+| | | | | | | |
+| wongnai | OpenThai | 500 | score | ??? | ??? | ??? |
+| wongnai | OpenThai | 500 | score | 0.6400 | 0.0424 | 3178.868 |
+| wongnai | Laya | 500 | score | 0.0340 | 0.0173 | 124.185 |
+| | | | | | | |
+| wisesight | OpenThai | 500 | choice | ??? | ??? | ??? |
+| wisesight | OpenThai | 500 | choice | 0.4960 | 0.3603 | 2000.341 |
+| wisesight | Laya | 500 | choice | 0.2500 | 0.2002 | 74.300 |
+| | | | | | | |
+| xlam_tools | OpenThai | 500 | choice | ??? | ??? | ??? |
+| xlam_tools | OpenThai | 500 | choice | 0.9920 | 0.0072 | 609.601 |
+| xlam_tools | Laya | 500 | choice | 0.8500 | 0.0598 | 45.588 |
+
+### OpenThaiONNX vs OpenThai Agreement (Cohen's kappa)
+
+#### CPU
+
+| Task | n | Kappa |
+|---|---:|---:|
+| **Overall** | **2704** | **1.000000** |
+| sib200_th / choice | 204 | 1.000000 |
+| wisesight / choice | 500 | 1.000000 |
+| wongnai / score | 500 | 1.000000 |
+| xlam_tools / choice | 500 | 1.000000 |
+| xnli_th / noul | 500 | 1.000000 |
+| xnli_th / choice | 500 | 1.000000 |
+
 ## Limitations
 
 - FP32 only.
-- CPU and CUDA execution providers only through the packaged client.
-- ONNX Runtime contrib operators make this artifact runtime-specific.
 - The model performs complete-request inference; generation caches are
   deliberately internal and reset for every call.
 - 262,144 tokens is an accepted graph boundary, not a performance guarantee.