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. | — | 0.6B/1.7B/4B/8B/14B/30B/32B/235B |
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. | — | $0.064DeepInfra |
Cheapest outputLowest output rate across every provider on OpenRouter, the lab included. May come from a different provider than the cheapest input. | — | $0.192DeepInfra |
| — | $30.00DeepInfra · fp8 | |
| Benchmarks | ||
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. | 75% | 70.7% |
AIME 2024Reasoning — single-answer accuracy on the 2024 invitational mathematics exam without tools | 91.82% | 85.7% |
AIME 2025Reasoning — single-answer accuracy on the 2025 invitational mathematics exam without tools | 83.48% | 81.5% |
GPQA DiamondScience — Graduate-level science questions in biology, physics, and chemistry — hard enough that subject-matter PhDs score around 65%. Higher is better. | 76.26% | 71.1% |
BullshitBench v2Nonsense detection — Given a confidently-worded but nonsensical prompt, does the AI spot that it makes no sense and push back — instead of playing along and inventing an answer? The score is how often it clearly called out the nonsense. Higher is better. | — | 6% |
IFEvalAgentic tasks — following verifiable instructions in generated responses | — | 83.4% |
MATH-500Reasoning — accuracy on the 500-problem MATH subset | — | 98% |
| Overview | ||
| Company | Mistral | Qwen |
| Release date | Sep 17 2025 | Apr 29 2025 |
| Access | Closed | Open Weight |
| Model details | View model | View model |
Other comparisons
Magistral Medium 1.2vsClaude Fable 5.2Qwen3vsClaude Fable 5.2Magistral Medium 1.2vsGPT-6.2 AstraQwen3vsGPT-6.2 AstraMagistral Medium 1.2vsGemini 4 NeonQwen3vsGemini 4 NeonMagistral Medium 1.2vsMuse EmberQwen3vsMuse EmberMagistral Medium 1.2vsGrok 4.7Qwen3vsGrok 4.7Magistral Medium 1.2vsDeepSeek-V4.1-Flash-1017Qwen3vsDeepSeek-V4.1-Flash-1017Frequently asked questions
Magistral Medium 1.2 leads Qwen3 on 4 of the 4 benchmarks they both report (LiveCodeBench, AIME 2024, AIME 2025, GPQA Diamond). Qwen3 shipped 141 days before Magistral Medium 1.2, so benchmark comparisons should account for the intervening progress.
Magistral Medium 1.2 is closed, while Qwen3 is open weight.
On LiveCodeBench, Magistral Medium 1.2 leads at 75% vs Qwen3 at 70.7%. On AIME 2024, Magistral Medium 1.2 leads at 91.82% vs Qwen3 at 85.7%. On AIME 2025, Magistral Medium 1.2 leads at 83.48% vs Qwen3 at 81.5%. On GPQA Diamond, Magistral Medium 1.2 leads at 76.26% vs Qwen3 at 71.1%.