‹ PublicAI Index
The LLM benchmark aggregator.
Mistral Large 3
Mistral
Strongest in Legal (#76 of 173), weakest in Safety (#261 of 329). Above par in 4 of 20 scopes. Among the models it meets almost everywhere, it finishes behind Claude Opus 4.7 and GPT-5.6 Sol and ahead of O3 Mini and GPT-5 nano.
§ 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.
Professional50.4−12#90/1941/2
Legal52.1−13.8#76/1731/1
Medical49.3−17.2#92/1491/1
Finance49.7−12.7#100/1621/1
Core abilities49.4−15.9#107/1991/4
General intelligence49.1−20.3#98/1781/3
Human preference58.1−10.7#115/3541/1
Human preference58.1−10.7#115/3541/1
Knowledge49.3−13.2#123/1971/3
Academic knowledge48.2−15.7#86/1231/1
Reasoning41.4−24#164/1952/5
Science44.7−18.8#92/1221/1
Mathematics43.8−20#112/1521/2
Reasoning36.4−32.6#137/1471/3
Coding39.7−26.5#168/1791/6
Agentic coding38.7−28.2#165/1731/5
Agents42.2−24.2#260/2982/6
Knowledge work37.5−33.2#165/2082/2
Safety46.4−15#261/3291/3
Factual grounding38.4−32.3#92/1041/1
§ 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 Large 3, left for the other.
§ 3 · Sources
Where the numbers come from
7 publications, 16 figures. Every one links to the page it was read from.
LMArena Text 1415
GDPval-AA 428
AA-Briefcase 229
Kagi LLM Benchmark 50.9%
SimpleBench 20.4%
Vals · CaseLaw 61.41%Vals · LegalBench 79.14%Vals · CorpFin 61.03%Vals · TaxEval 73.06%Vals · MortgageTax 52.11%Vals · MedQA 82.23%Vals · SWE-bench Verified 41.4% (Mini-SWE-agent)Vals · GPQA Diamond 68.43%Vals · MMLU Pro 79.82%Vals · AIME 42.92%
Vectara · Factual consistency 85.5%
Badge
[](https://publicai.io/model-index/m/mistral-large-3)