Compare AI models
| Specifications | ||
Context windowHow much information the model can work with in one request, measured in tokens — pieces of words or other data. This includes your question, attachments, conversation and reply. Limits can vary by model variant and provider. | 128k | — |
| API pricingUSD per 1M tokens · lower wins | ||
Cheapest inputLowest input rate across every provider on OpenRouter, the lab included. The cheapest endpoint may run a quantised build or a shorter context — see "Available from" on the model page. | $0.40Mistral | — |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $2.00Mistral | — |
| $180Mistral · zdr | — | |
| Benchmarks | ||
HumanEvalFunction synthesis — 164 small Python problems: the AI is given a function's description and has to write the working function. This was the coding benchmark of the GPT-3.5 and GPT-4 era, before the field moved to fixing real bugs in real repositories. Higher is better. | 92.1% | 85.7% |
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 77.2% | 72% |
MMMUMultimodal — Tests the AI on understanding images and text together across many college subjects. Higher is better. | 66.1% | 63.2% |
LiveCodeBenchCompetitive coding — Coding problems published so recently the AI can't have seen them in training — a contamination-free test of raw programming skill. Higher is better. | 30.3% | — |
IFEvalAgentic tasks — following verifiable instructions in generated responses | 89.4% | — |
MATHReasoning — accuracy on the original competition mathematics test set | — | 73% |
MATH-500Reasoning — accuracy on the 500-problem MATH subset | 91% | — |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 57.1% | — |
MMLUGeneral knowledge — A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better. | — | 86.2% |
MathVistaMultimodal reasoning — mathematical reasoning about visual information | — | 68.1% |
| Overview | ||
| Company | Mistral | SpaceXSI |
| Release date | May 7 2025 | Aug 13 2024 |
| Access | Closed | Closed |
| Model details | View model | View model |
Other comparisons
Mistral Medium 3vsClaude Fable 5.2Grok‑2 minivsClaude Fable 5.2Mistral Medium 3vsGPT-6.2 AstraGrok‑2 minivsGPT-6.2 AstraMistral Medium 3vsGemini 4 HydrogenGrok‑2 minivsGemini 4 HydrogenMistral Medium 3vsMuse EmberGrok‑2 minivsMuse EmberMistral Medium 3vsDeepSeek-V4.1-Flash-1017Grok‑2 minivsDeepSeek-V4.1-Flash-1017Mistral Medium 3vsKimi K3.1 ProGrok‑2 minivsKimi K3.1 ProFrequently asked questions
Mistral Medium 3 leads Grok‑2 mini on 3 of the 3 benchmarks they both report (HumanEval, MMLU-Pro, MMMU). Grok‑2 mini shipped 267 days before Mistral Medium 3, so benchmark comparisons should account for the intervening progress.
Published specifications for these two models are limited — see each model page for the latest details.
On HumanEval, Mistral Medium 3 leads at 92.1% vs Grok‑2 mini at 85.7%. On MMLU-Pro, Mistral Medium 3 leads at 77.2% vs Grok‑2 mini at 72%. On MMMU, Mistral Medium 3 leads at 66.1% vs Grok‑2 mini at 63.2%.