LuminaBench
mistralai / Published benchmarks

Mistral Medium 3.5

Benchmark results and API pricing.

Best recorded results

Benchmark scores

Best eligible score in each benchmark.

Legal Research Bench887High · ExperimentalMedCode933High · ExperimentalMedScribe926High · ExperimentalProofBench v1.1890High · ExperimentalTax Agent Bench880High · Experimental

API pricing

USD / 1M tokens

Verified pricing is unavailable for this model.

Legal Research Bench

Legal Research Bench

Version 1-updated-2026-11-05
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

887Best score · High
Explore Legal Research Bench
MedCode

MedCode

Version 1-updated-2026-10-29
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

933Best score · High
Explore MedCode
MedScribe

MedScribe

Version 1-updated-2026-10-29
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

926Best score · High
Explore MedScribe
ProofBench v1.1

ProofBench v1.1

Version 1.1-updated-2026-10-29
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

890Best score · High
Explore ProofBench v1.1
Tax Agent Bench

Tax Agent Bench

Version 1-updated-2026-11-05
Lumina VerifiedLumina checked the published sources, versions, model labels, score extraction and conversion. External evaluations were run by their publishers; Lumina has not independently rerun or reproduced them.

1 configuration · Publisher-run · Experimental conversion

880Best score · High
Explore Tax Agent Bench