NEWMEN
← back to all models

Mistral Large

mistralai/mistral-large

Mistralopen-weight128K context1 providerIntelligence 70.8

Cheapest provider

$2.00 / 1M

Mistral

Fastest provider (p95)

74 tok/s

Mistral

Intelligence (composite)

70.8

MMLU-Pro · HumanEval · math · GPQA

Per-provider performance

Latency / throughput / uptime / price measured across providers over the last 30 minutes of live traffic. This is what proves “sourced cheapest” — Atlas mode draws on these per call to serve the cheapest path that holds quality.

ProviderQuantInput $/1MOutput $/1MLatency p50 / p95Throughput p50 / p95Uptime 30mSuccess
Mistralundisclosed$2.0000$6.0000490ms / 5035ms38 / 74 tok/s99.49%97.5%

“—” means live telemetry hasn’t accumulated enough recent traffic for that endpoint. “undisclosed” means the provider serves the model but doesn’t expose the quantization label (typically running fp8 / int8 internally).

Intelligence breakdown

Composite score is a weighted average of public benchmarks (30% MMLU-Pro, 25% code pass@1, 25% math, 20% GPQA). Numbers come from model cards and the Artificial Analysis intelligence harness; missing components are renormalised over what’s present.

MMLU-Pro

70.4

broad reasoning

Code

84.7

pass@1 (HumanEval / LiveCodeBench)

MATH

73.8

math accuracy

GPQA Diamond

50.1

hard reasoning

Source: Mistral Large (Nov 2024 / 2411 model card)

How Atlas mode sources Mistral Large

  • Strict mode — pin Mistral Large exactly and we pass it straight through, sourced from the cheapest provider above. The same model, no substitutions — currently Mistral at $2.00/1M.
  • Atlas mode — the default. Each call is auto-optimized for the cheapest path that holds quality, at least 5% off going direct from call one and climbing as it ramps. You always see which model served the call and exactly what you saved — thumbs-down anything you don’t like for a full refund.