---
base_model: CohereLabs/BLS-Mini-Code-1.0
pipeline_tag: text-generation
library_name: gguf
tags:
  - cohere
  - moe
  - code
  - reasoning
  - gguf
quantized_by: Arki05
---

# BLS-Mini-Code-1.0 — GGUF

GGUF quantizations of [CohereLabs/BLS-Mini-Code-1.0](https://huggingface.co/CohereLabs/BLS-Mini-Code-1.0),
a 30.5B-total / ~2.9B-active sparse MoE code model by Cohere (`cohere2moe`
architecture: Command-R7B-style hybrid SWA/full attention with NoPE on global
layers, parallel residual blocks, 128 fine-grained experts with sigmoid top-8
routing, reasoning-by-default chat format).

> **Status / requirements:** needs llama.cpp with `cohere2moe` support —
> [PR #24260](https://github.com/ggml-org/llama.cpp/pull/24260) (not yet merged).
> Build that branch until it lands. The upstream model repo currently ships
> **no license**; these files inherit whatever terms Cohere attaches to the
> original weights.

## Quants

All quality numbers are measured against the **bf16 model as ground truth**.
The headline table uses **wikitext-2 (test)** — the only evaluation set that is
fully held out from the imatrix calibration data — plus HumanEval/HumanEval+
(pass@1, greedy, thinking on, 6k token budget; remaining quants in progress).

| file | size | PPL | mean KLD | top-1 % | HumanEval | HumanEval+ |
|---|---|---|---|---|---|---|
| BF16 (2 shards) | 61.0 GB | 7.7126 | — | — | | |
| Q8_0 | 32.4 GB | 7.7356 | 0.007010 | 96.458 | 92.07 | 89.02 |
| Q6_K | 25.1 GB | 7.7558 | 0.015611 | 94.602 | 93.29 | 88.41 |
| Q5_K_M | 21.7 GB | 7.8333 | 0.020963 | 93.811 | 95.73 | 92.68 |
| Q4_K_M | 18.6 GB | 7.9468 | 0.041855 | 91.342 | 93.29 | 90.24 |
| IQ4_XS | 16.4 GB | 7.9794 | 0.049137 | 90.705 | 92.68 | 88.41 |
| IQ3_M | 13.6 GB | 8.2776 | 0.112035 | 85.919 | 90.85 | 87.20 |
| IQ2_M | 10.3 GB | 9.9756 | 0.283656 | 77.616 | 84.15 | 79.88 |
| IQ2_XS | 9.2 GB | 11.0666 | 0.426120 | 73.339 | 79.88 | 77.44 |
| IQ2_XXS | 8.3 GB | 12.6780 | 0.549859 | 69.743 | 59.15 | 59.15 |

HumanEval is pass@1 over 164 problems, so single-token greedy flips on a handful
of problems move the score by a few points - read it as a sanity check, not a
fine-grained ranking. The Q4-through-Q8 quants are statistically interchangeable
on it (the spread is noise); **mean KLD and top-1 % are the reliable quality
ordering**. The slope only becomes clear lower down: IQ3_M holds up, the IQ2 tier
degrades visibly, and IQ2_XXS falls off a cliff (identical HumanEval/HumanEval+
is the giveaway - it produces enough malformed code that the extra tests prune
almost nothing further).

Recommendations: **Q5_K_M** if you have the memory (effectively lossless),
**IQ4_XS** for the best size/quality ratio (matches Q4_K_M at -2.2 GB),
**IQ3_M** as the smallest quant still reasonable for code. The IQ2 tier exists
for memory-constrained setups and degrades noticeably - use with expectations
set accordingly. Embeddings are tied (also the output head) and kept at q6_K
on Q4_K_M and below.

## Per-domain breakdown

The three sets below are also part of the imatrix calibration corpus, so their
numbers carry a mild in-distribution bias - read them as domain comparisons
rather than held-out scores. All corpora are included in
`eval-corpora.tar.zst` for reproduction.

### General / multilingual (calibration_datav3)

[bartowski's calibration_datav3](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8):
the de-facto community calibration mix - short English prose, multilingual
snippets, code fragments, technical text and deliberate noise sections
(~275 kB).

| file | PPL | mean KLD | top-1 % |
|---|---|---|---|
| BF16 | 9.0079 | — | — |
| Q8_0 | 9.0261 | 0.008424 | 96.788 |
| Q6_K | 9.0351 | 0.014500 | 95.286 |
| Q5_K_M | 9.0491 | 0.019470 | 94.506 |
| Q4_K_M | 9.1607 | 0.036786 | 92.031 |
| IQ4_XS | 9.1125 | 0.039540 | 91.882 |
| IQ3_M | 9.4710 | 0.087992 | 87.714 |
| IQ2_M | 10.2735 | 0.208782 | 80.580 |
| IQ2_XS | 11.1268 | 0.319906 | 76.376 |
| IQ2_XXS | 12.3083 | 0.427367 | 72.173 |

### Code

A seeded random sample of real source files from the
[llama.cpp](https://github.com/ggml-org/llama.cpp) tree (MIT): C/C++ core and
ggml, Python conversion tooling, shell scripts; capped at 25 kB per file,
~400 kB total. Note how confident the model is on code (PPL ~2.4) - and that
top-1 agreement holds up better here than on prose at every quant level.

| file | PPL | mean KLD | top-1 % |
|---|---|---|---|
| BF16 | 2.4043 | — | — |
| Q8_0 | 2.4108 | 0.005231 | 98.512 |
| Q6_K | 2.4123 | 0.008321 | 97.731 |
| Q5_K_M | 2.4155 | 0.012198 | 97.145 |
| Q4_K_M | 2.4314 | 0.025947 | 95.898 |
| IQ4_XS | 2.4452 | 0.030205 | 95.472 |
| IQ3_M | 2.4996 | 0.072891 | 92.991 |
| IQ2_M | 2.7561 | 0.186894 | 88.646 |
| IQ2_XS | 3.0247 | 0.290555 | 85.260 |
| IQ2_XXS | 3.2342 | 0.368478 | 83.263 |

### Chat (model-native format)

Hand-written for this release: 13 short programming conversations
(Python/SQL/C/Rust/git topics, two in German), each with a thinking block,
plus one complete tool-call round trip - rendered in the model's raw turn-token
dialect (`<|START_OF_TURN_TOKEN|>`, `<|START_THINKING|>`, `<|START_ACTION|>`,
...). This exercises the control-token and expert-routing paths that real chat
traffic hits and plain text never does. Small set (~7 chunks) - treat the
numbers as indicative.

| file | PPL | mean KLD | top-1 % |
|---|---|---|---|
| BF16 | 1.9660 | — | — |
| Q8_0 | 1.9866 | 0.022651 | 98.431 |
| Q6_K | 1.9906 | 0.031189 | 98.170 |
| Q5_K_M | 1.9820 | 0.025972 | 97.778 |
| Q4_K_M | 1.9641 | 0.070232 | 96.993 |
| IQ4_XS | 1.9866 | 0.058722 | 96.601 |
| IQ3_M | 2.0809 | 0.081966 | 94.902 |
| IQ2_M | 2.1412 | 0.173477 | 92.288 |
| IQ2_XS | 2.1742 | 0.251918 | 89.412 |
| IQ2_XXS | 2.2247 | 0.297151 | 87.974 |

## Reasoning / chat template

These GGUFs embed an **additively normalized** chat template (also in this repo
as `chat_template.jinja`): the standard `enable_thinking` /
`reasoning_content` conventions are mapped onto Cohere's native `reasoning` /
`reasoning_effort` / `thinking` variables, so llama.cpp detects reasoning
support automatically (`thinking = 1`), separates `reasoning_content` from
`content`, and supports thinking toggles. All Cohere-native variables keep
working; rendering is byte-identical for native invocations.

```bash
llama-server -m BLS-Mini-Code-1.0-Q5_K_M.gguf --jinja
```

- thinking on (default): response arrives as `reasoning_content` + `content`
- disable thinking per request: `"chat_template_kwargs": {"enable_thinking": false}`
  (or Cohere-native: `{"reasoning_effort": "none"}`)
- tool calling works through the OpenAI-compatible API (parallel calls included)

## imatrix

`BLS-Mini-Code-1.0.imatrix` (included) was computed on the **bf16** model over
the v3 + code + chat mix described above (326x512-token chunks), reaching full
coverage of all 128 experts in every layer.

## Validation

- f32 logit-level parity vs HF transformers on a truncated-expert variant of
  the checkpoint (full-vocab comparison at every position): top-1 agreement
  26/27, mean |dlogprob| 0.012 - the only disagreement a 0.013 near-tie.
- Tool calling, parallel calls, multi-turn with reasoning passback, and a live
  agentic tool-execution loop verified end to end via `llama-server`.
- 500k context advertised by the model; KV cache at long context stays small
  thanks to iSWA (only 13 of 49 layers are global; ~13.6 GB KV at 500k).