‹ PublicAI Index
The LLM benchmark aggregator.
Mistral Medium 3.5
Mistral
Strongest in Human preference (#99 of 354), weakest in Agents (#191 of 298). Above par in 2 of 17 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of Command A and Llama 3.3 70B Instruct.
§ 1 · Profile
What it is good at
Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.
Human preference59.3−9.5#99/3541/1
Human preference59.3−9.5#99/3541/1
Knowledge46.9−15.6#142/1971/3
Academic knowledge42.6−21.3#101/1231/1
Coding39.5−26.7#172/1791/6
Agentic coding39−27.9#163/1731/5
Core abilities41.9−23.4#173/1992/4
General intelligence39.6−29.8#152/1782/3
Professional38.7−23.7#184/1941/2
Medical39.6−26.9#133/1491/1
Finance39.7−22.7#140/1621/1
Legal32.4−33.5#173/1731/1
Reasoning35.3−30.1#187/1951/5
Science26.1−37.4#117/1221/1
Mathematics36.4−27.4#135/1521/2
Agents46.5−19.9#191/2982/6
Knowledge work44.4−26.3#127/2082/2
§ 2 · Head to head
What it beats, and what beats it
The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Mistral Medium 3.5, left for the other.
§ 3 · Sources
Where the numbers come from
6 publications, 19 figures. Every one links to the page it was read from.
LMArena Text 1427
Artificial Analysis Intelligence Index 5
GDPval-AA 769
AA-Briefcase 523
Kagi LLM Benchmark 41.4%
Vals · Legal Research Bench 9.13%Vals · CaseLaw 44.16%Vals · Finance Agent 32.1%Vals · CorpFin 58.78%Vals · TaxEval 67.99%Vals · MortgageTax 28.89%Vals · MedCode 33.75%Vals · MedScribe 67.73%Vals · SWE-bench Verified 66.4% (Mini-SWE-agent)Vals · Vibe Code Bench 2.89% (OpenHands)Vals · Code Migration 5.13%Vals · GPQA Diamond 34.85%Vals · MMLU Pro 75.33%Vals · ProofBench 9%
Badge
[](https://publicai.io/model-index/m/mistral-medium-3-5)