Compare / Mistral Small 3 vs Mistral Small 4

Mistral Small 3 vs Mistral Small 4

Which is cheaper, which is more capable, and which does better at each kind of work, from 9 results on 2 sources that measured both models.

Results as of 8 October 2026 (catalogue release 2026-10-08-4199ce0620fe).

Mistral Small 3
Mistral AI · profile
Mistral Small 4
Mistral AI · profile
Results compared
9
Shared sources
2

Mistral Small 3 or Mistral Small 4? The short answer

Mistral Small 4 is the newer version and scores higher (Intelligence Index 11.3 against 6.7), but Mistral Small 3 is cheaper (4.6 times less per token): upgrade if the gain is worth the price.

  • Mistral Small 3 is cheaper: $0.058 against $0.26 per million tokens, 4.6 times less.
  • Mistral Small 4 scores higher on the Artificial Analysis Intelligence Index: 11.3 against 6.7.
  • Mistral Small 4 is newer: released 16 Mar 2026, against 30 Jan 2025 for Mistral Small 3.
  • On third-party benchmarks, on the results both models have: Mistral Small 4 does better for reasoning and knowledge and writing.

At a glance

Mistral Small 3Mistral Small 4
Price per million tokens$0.050 in · $0.080 out$0.15 in · $0.60 out
Blended (3 in : 1 out)$0.058$0.26
Cheapest host, blended–$0.26 (Mistral)
Cached input, per million–$0.015 cached input
Intelligence Index6.711.3
Spring Prompt overall––
Output speed, tokens a second–164
Context window32,768 tokens262,144 tokens
Released30 Jan 202516 Mar 2026 (newer)
WeightsOpenOpen
Developer based inFranceFrance

Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.

Which is better for what

Each model's rank among models available today, for each kind of work, from our own benchmarks and licensed sources. The better rank is in green.

Use caseMistral Small 3Mistral Small 4
Agents and tool useMulti-step tasks with tools: support desks, coding agents, function calls not measured 132 of 164
Reasoning and knowledgeHard questions across science, maths and general knowledge 193 of 198 131 of 198
WritingWhat people prefer in blind comparisons, and judged writing quality 156 of 188 116 of 188
Sticking to the factsSummarising without inventing things, and factual answers 9 of 152 not measured
Speed under pressureGood decisions against a real clock (fast chess) not measured 9 of 22

Every shared result

Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.

MeasureMistral Small 3Mistral Small 4
Artificial Analysis reported by Artificial Analysis · 3 results
Artificial Analysis Intelligence Indexindex score, higher is better 6.7 11.3
GPQA Diamond% of questions, higher is better 46.2% 76.9%
Humanity's Last Exam% of questions, higher is better 3.8% 9.9%
UGI Leaderboard reported by UGI Leaderboard · 3 results
Writing score · UGIscore out of 100, higher is better 33.9 40.3no reasoning
Requested-length error · UGI% off the requested word count, lower is better 8.0% 22.0%no reasoning
Style adherence · UGIscore from 0 to 1, higher is better 0.35 0.31high reasoning
Show all 9 results (3 more)
MeasureMistral Small 3Mistral Small 4
Artificial Analysis reported by Artificial Analysis · 3 results
IFBench% of instructions, higher is better 26.4% 48.2%
Long-context reasoning (AA-LCR)% of questions, higher is better 0.0% 49.7%
τ²-bench telecom% of tasks, higher is better 19.6% 41.2%

Quick answers

Is Mistral Small 3 better than Mistral Small 4?

Mistral Small 4 is the newer version and scores higher (Intelligence Index 11.3 against 6.7), but Mistral Small 3 is cheaper (4.6 times less per token): upgrade if the gain is worth the price.

Which is cheaper, Mistral Small 3 or Mistral Small 4?

Mistral Small 3 is cheaper: $0.058 against $0.26 per million tokens, 4.6 times less.

Run Mistral Small 3 and Mistral Small 4 on your own prompt

Benchmarks aren't your data. Try both side by side in the playground, with the cost of every answer.