Model

Mistral Medium 3

CLOSED

Cost-efficient closed frontier model claimed at ~90% of Claude 3.7 Sonnet across benchmarks, 8x cheaper.

ReasoningCodingVisionFunction callingTool useAgentic

1220.2

Elo · rank #178

Parameters

Undisclosed

Active params

—

Context

128K tokens

Architecture

Undisclosed dense transformer, multimodal (closed weights)

License

Proprietary

Languages

—

API price (in/out)

$0.4 / $2

Modalities

text · vision

Benchmark results

Bar shows position within the tracked field; marker = field best

LiveCodeBenchCoding44.8%#126

best: DeepSeek-V4-Pro (Think Max) · 93.5%

MMLU-ProKnowledge75.3%#75

best: Claude Fable 5 · 91.5%

MMMUVision63.0%#71

best: Claude Fable 5 · 89.3%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

$0.4

Output / M tok

Mistral Medium family

Elo progression across releases

Resources

API price $0.4/$2 · each benchmark row carries its own source badge (see methodology)