Known-issues sheets for local language models Rev 0.1.0 · reviewed every 7 days

Model datasheet · Mistral Small

Mistral Small 4

Mistral folded Magistral, Pixtral and Devstral into one model: reasoning, vision and agentic coding, Apache licensed, 6.5B active of 119B. The catch is attention. It kept full multi-head attention where everything else its size uses grouped-query, so its KV cache costs 576KB per token — the most expensive on this site, and more than a 358B model's. Cheap to compute, dear to give context to.

Vendor
Mistral AI
Architecture
MoE · 119B total / ~6.5B active
Context
262,144 tokens
License
Apache-2.0
Released
2026-03-16
Vision
Yes

→ GGUF (llama.cpp / LM Studio / Ollama)

§1 Characteristics by quantization band

Q2–Q3 — 2–3 bit

54.4GB weights · e.g. Q3_K_M + KV: 4.5GB at 8K · 18GB at 32K · 72GB at 128K
Coding 5/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q3_K_M · aggregated · 2026-08-31

    Editorial derivation: two points down. At 3-bit a sparse MoE loses expert-routing precision first, which surfaces as dropped tool arguments and broken multi-file edits before prose visibly suffers.

Q4–Q5 — 4–5 bit

73.8GB weights · e.g. Q4_K_M + KV: 4.5GB at 8K · 18GB at 32K · 72GB at 128K
Math & reasoning 7/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q4_K_M · aggregated · 2026-08-31

    Editorial derivation: one point down from the measured bf16 figure. 4-bit is the band a model this size is actually run at, and the loss is real but modest.

Coding 6/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q4_K_M · aggregated · 2026-08-31

    Editorial derivation: one point down from the measured bf16 figure. 4-bit is the band a model this size is actually run at, and the loss is real but modest.

Q6–Q8 — 6–8 bit

99.4GB weights · e.g. Q6_K + KV: 4.5GB at 8K · 18GB at 32K · 72GB at 128K
Math & reasoning 8/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q6_K · aggregated · 2026-08-31

    Editorial derivation: no drop from the measured bf16 figure. 6-8 bit is effectively lossless on a model this size.

Coding 7/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q6_K · aggregated · 2026-08-31

    Editorial derivation: no drop from the measured bf16 figure. 6-8 bit is effectively lossless on a model this size.

Factuality 7/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q6_K · aggregated · 2026-08-31

    Editorial derivation: no drop from the measured bf16 figure. 6-8 bit is effectively lossless on a model this size.

Long context 7/10
data checked Aug 2026
  • quant-degradation-community — editorial band derivation @ Q6_K · aggregated · 2026-08-31

    Editorial derivation: no drop from the measured bf16 figure. 6-8 bit is effectively lossless on a model this size.

FP16 — full precision (fp16/bf16)

238GB weights · e.g. bf16 + KV: 4.5GB at 8K · 18GB at 32K · 72GB at 128K
Math & reasoning 8/10
data checked Aug 2026
  • vendor-model-cardAIME 2025 = 83.8 @ bf16 · vendor-reported · 2026-08-31

    GPQA Diamond 71.2. Reasoning is toggled per request with reasoning_effort, so the figure assumes it is on.

Coding 7/10
data checked Aug 2026
  • vendor-model-cardLiveCodeBench = 63.6 @ bf16 · vendor-reported · 2026-08-31

    Mistral claim it beats gpt-oss-120b here while producing 20% less output. No SWE-bench figure is published, which is the number that would matter most for the agentic coding it absorbed from Devstral.

Factuality 7/10
data checked Aug 2026
  • vendor-model-cardMMLU-Pro = 78 @ bf16 · vendor-reported · 2026-08-31

    MMMU-Pro 60 on the vision side. No hallucination-rate benchmark is published, so this is a knowledge score rather than a truthfulness one.

Long context 7/10
data checked Aug 2026
  • artificial-analysis — AA-LCR long-context reasoning (link-only — no numbers republished) @ bf16 · aggregated · 2026-08-31

    Cited by link under their terms. Mistral report it as matching gpt-oss-120b. Worth reading against this model's KV cost: the context is affordable to reason over and expensive to hold.

§2 Known issues & what fixes them

ISSUE-01 severe Q2–Q3 / Q4–Q5 / Q6–Q8 / FP16 · long-context

The most expensive KV cache in this catalogue, and by a long way. Its config keeps full multi-head attention — 32 key-value heads against 32 attention heads — where every comparable model uses grouped-query attention to shrink exactly this. The result is 576KB per token: 18GB of cache at 32K context, about 72GB at 128K, on top of 74GB of 4-bit weights. A 358B model with grouped-query attention costs less. The sparsity that makes it fast to run does nothing for the memory it needs to hold a conversation.

No fix
No known fix — choose a different model

Architectural. Budget the cache explicitly — the RUNS AT line prices it at the context you pick — or run it at short context, which is where its 6.5B active parameters pay off.

Evidence · 2 sources · community-consensus
ISSUE-02 moderate Q2–Q3 / Q4–Q5 / Q6–Q8 / FP16 · tool-calling, instruction-following

Sold as the model that absorbed Devstral's agentic coding, and published with no tool-calling or instruction-following benchmark to show for it — no BFCL, no IFEval, no SWE-bench. Those cells are blank here rather than estimated, which the coverage discount then penalises. Every figure that does exist is Mistral's own; no independent replication was found.

No fix
No known fix — choose a different model

A gap in the evidence, not a defect in the model. Treat the agentic claim as unverified until someone benchmarks it.

Evidence · 2 sources · community-consensus