Compare AI models
| Specifications | ||
ParametersThe learned values that shape how a model responds. B means billion and T means trillion. For mixture-of-experts models, the total count is larger than the number active for each token. Parameter count alone does not measure quality. | — | 0.5B/1.5B/3B/7B/14B/32B/72B |
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 | 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. | $0.44Mistral | $0.288DeepInfra |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | $2.20Mistral | $0.32DeepInfra |
| $198Mistral | $92.00DeepInfra · fp8 | |
| Benchmarks | ||
MMLU-ProKnowledge — accuracy on the ten-option MMLU-Pro benchmark | 76.8% | 71.1% |
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. | — | 55.5% |
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. | — | 86.6% |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 84.1% |
MATHReasoning — accuracy on the original competition mathematics test set | — | 83.1% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 65.7% | — |
IFBenchInstruction following — Tests whether the AI can follow detailed instructions and satisfy multiple constraints at once. Higher is better. | 40.8% | — |
GSM8KGrade-school math — Grade-school maths word problems that take a few steps of arithmetic to work through. It separated the models of 2022 and 2023 sharply, then saturated. One caveat on the historical numbers: OpenAI included part of the GSM8K training set in GPT-4's pre-training mix, so GPT-4's score is not a clean few-shot result. Higher is better. | — | 95.8% |
MMMU-ProMultimodal reasoning — A tougher version of MMMU — college-level questions that mix images, diagrams, and text together. Higher is better. | 43.8% | — |
| Overview | ||
| Company | Mistral | Qwen |
| Release date | Aug 12 2025 | Sep 19 2024 |
| Access | Closed | Open Weight |
| Model details | View model | View model |
Other comparisons
Mistral Medium 3.1vsClaude Fable 5.2Qwen2.5vsClaude Fable 5.2Mistral Medium 3.1vsGPT-6.2 AstraQwen2.5vsGPT-6.2 AstraMistral Medium 3.1vsGemini 4 NeonQwen2.5vsGemini 4 NeonMistral Medium 3.1vsMuse EmberQwen2.5vsMuse EmberMistral Medium 3.1vsGrok 4.7Qwen2.5vsGrok 4.7Mistral Medium 3.1vsDeepSeek-V4.1-Flash-1017Qwen2.5vsDeepSeek-V4.1-Flash-1017Frequently asked questions
Mistral Medium 3.1 leads Qwen2.5 on 1 of the 1 benchmark they both report (MMLU-Pro). Qwen2.5 shipped 327 days before Mistral Medium 3.1, so benchmark comparisons should account for the intervening progress.
Context windows are 128k (Mistral Medium 3.1) vs 128k (Qwen2.5). Mistral Medium 3.1 is closed, while Qwen2.5 is open weight.
On MMLU-Pro, Mistral Medium 3.1 leads at 76.8% vs Qwen2.5 at 71.1%.