Mistral Small 4

Mistral Small 4 is a non-reasoning model from Mistral. 15 benchmarks count toward its score, in 6 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.38.4 ±5.7
CoverageShare of the index weight with results.85%
SpeedOutput tokens per second.166/s
Input / 1MUS dollars per 1M input tokens.$0.15 batch $0.075
Output / 1MUS dollars per 1M output tokens.$0.6 batch $0.3 US dollars per 1M output tokens in a batch.
ContextMaximum tokens in one request.256K
EloLMArena rating and rank.N/A

50 is the middle of the board. The range shows the doubt in the index. Batch work costs less.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
34.2
CodingCode writing and repair.
42.2
ReasoningLogic problems and puzzles.
40.8
MultimodalTasks with images and text.
36.5
KnowledgeFacts and expert knowledge.
40.0
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
38.0
MathMath problems.
N/A

Results

15 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
Artificial Analysis GPQA DiamondKnowledge76.9%47.4Artificial Analysis
Artificial Analysis MMMU-ProMultimodal56.8%36.5Artificial Analysis
Artificial Analysis Long Context ReasoningReasoning49.7%42.6Artificial Analysis
Artificial Analysis IFBenchInstruction48.2%38.0Artificial Analysis
τ²-Bench Tool-Agent-User EvaluationAgentic41.2%28.4Victor Barres et al.
Artificial Analysis SciCodeCoding38.8%46.7Artificial Analysis
Artificial Analysis Coding IndexCoding26.6%37.7Artificial Analysis
Artificial Analysis Omniscience AccuracyKnowledge21.7%40.7Artificial Analysis
Terminal-Bench 2.0Agentic16.9%34.74 Jun 2026Vals AI
Artificial Analysis Intelligence IndexKnowledge11.3%36.5Artificial Analysis
Artificial Analysis Humanity's Last ExamKnowledge9.9%35.3Artificial Analysis
Artificial Analysis Agentic IndexAgentic1.4%38.3Artificial Analysis
Critical Physics TasksReasoning0.3%39.0Artificial Analysis
GDPval-AA normalizedAgentic0.0%35.6Artificial Analysis
Vibe Code Bench v1.1Coding0.0%42.2OpenHands21 Sept 2026Vals AI

15 benchmarks count, from 15 of 15 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsVals AI, collected directlyNo licence stated. Read from the public leaderboard and credited to Vals AI

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from Mistral

Mistral Medium 3.5 128B49.8Mistral Medium 3.543.0Mistral Small 4 (Reasoning)38.5Magistral Medium38.3Magistral Small35.7