Model comparison
GPT-5 Nano vs Mistral Medium
Mistral Medium is the stronger model overall, scoring 36.0 to 33.4 on the Noometry Index. GPT-5 Nano costs 22× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Last verified . 28 shared benchmarks.
Mistral AISummary
- They share 28 benchmarks with published results for both. GPT-5 Nano scores higher in 3 categories and Mistral Medium in 7 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Medium leads 60.2 to 39.4.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 81.1% for GPT-5 Nano and 32.2% for Mistral Medium.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- GPT-5 Nano accepts more context: 400K tokens versus 262K.
- Mistral Medium has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | Mistral Medium | |
|---|---|---|
| Provider | OpenAI | Mistral AI |
| Noometry Index | 33.4 | 36.0 |
| Released | 2025-08-07 | 2023-12-11 |
| Weights | Proprietary | Open |
| Context window | 400K | 262K |
| Max output | 128K | 262K |
| Input $ / M tokens | $0.05 | $1.50 |
| Output $ / M tokens | $0.40 | $7.50 |
| Results tracked | 48 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-5 Nano: 33.3 (#248), Mistral Medium: 33.8 (#241)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| WeirdML | 38.1% | 43.7% |
| LMArena Coding | 1351 | 1434 |
| ALE-Bench | 718.67 | 763.98 |
| FrontierCode | — | 8% |
| SWE-bench Verified (bash only) | 34.8% | — |
| SciCode | — | 40.2% |
Agentic & Tool Use Mistral Medium leads
GPT-5 Nano: 25.6 (#99), Mistral Medium: 28.1 (#84)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| Berkeley Function Calling Leaderboard | 51.5% | 37.7% |
| Terminal-Bench | 21.8% | — |
Reasoning Mistral Medium leads
GPT-5 Nano: 16.4 (#291), Mistral Medium: 23.3 (#165)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 62.2% | 50% |
| LMArena Hard Prompts | 1329 | 1426 |
| DTBench | 62.7% | 75.5% |
| LMCA | 7.9% | 26.1% |
| ARC-AGI-2 | 2.6% | — |
| ARC-AGI-1 | 20.7% | — |
| CritPt | — | 0% |
| Chess Puzzles | 27% | — |
| Mystery Game Puzzles | 9% | — |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 139.39 | — |
| ForecastBench | 59.1 | — |
Math GPT-5 Nano leads
GPT-5 Nano: 29.1 (#221), Mistral Medium: 27.2 (#238)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.1% | 32.2% |
| ProofBench | 12% | 9% |
| LMArena Math | 1318 | 1409 |
| MATH Level 5 | 95.2% | 81.6% |
| FrontierMath (Feb 2025 set) | 8.3% | 0.3% |
| FrontierMath (Tiers 1-3) | 20% | — |
| FrontierMath Tier 4 | 2.4% | — |
| Omni-MATH | 54.6% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-5 Nano leads
GPT-5 Nano: 35.8 (#166), Mistral Medium: 24.9 (#260)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| GPQA Diamond | 69.4% | 59.5% |
| Vectara Hallucination Rate | 10.5% | 22.7% |
| LMArena Expert | 1321 | 1408 |
| Humanity's Last Exam | — | 4.5% |
| SimpleQA Verified | 11.7% | — |
| MMLU-Pro | 77.8% | — |
| GPQA (HELM) | 67.9% | — |
Multimodal Mistral Medium leads
GPT-5 Nano: 30.7 (#102), Mistral Medium: 34.6 (#82)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| LMArena Vision | 1159 | 1172 |
| VPCT | 37.2% | — |
Multilingual Mistral Medium leads
GPT-5 Nano: 45.5 (#167), Mistral Medium: 52.3 (#86)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1313 | 1408 |
| LMArena Chinese | 1355 | 1448 |
| LMArena German | 1329 | 1433 |
| LMArena Korean | 1269 | 1382 |
| LMArena Russian | 1296 | 1411 |
| LMArena Spanish | 1360 | 1433 |
| LMArena French | — | 1460 |
| LMArena Japanese | — | 1378 |
Instruction Following GPT-5 Nano leads
GPT-5 Nano: 75.2 (#72), Mistral Medium: 73.9 (#111)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1306 | 1398 |
| IFEval | 93.2% | — |
Long Context Mistral Medium leads
GPT-5 Nano: 31.3 (#275), Mistral Medium: 42.9 (#108)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1313 | 1406 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference Mistral Medium leads
GPT-5 Nano: 39.4 (#244), Mistral Medium: 60.2 (#97)
| Benchmark | GPT-5 Nano | Mistral Medium |
|---|---|---|
| LMArena Text | 1320 | 1424 |
| LMArena Creative Writing | 1250 | 1391 |
| LMArena Multi-Turn | 1311 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| EQ-Bench Creative Writing | 705 | — |
| WildBench | 80.6% | — |
Frequently asked questions
Is GPT-5 Nano better than Mistral Medium?
Mistral Medium is the stronger model overall, scoring 36.0 to 33.4 on the Noometry Index. GPT-5 Nano costs 22× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or Mistral Medium?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is GPT-5 Nano or Mistral Medium better for coding?
They score almost the same on coding (33.3 vs 33.8); test both on your own repository before choosing.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 262K.
How many benchmarks do GPT-5 Nano and Mistral Medium share?
28 benchmarks have published results for both models. GPT-5 Nano has 48 scored results on Noometry and Mistral Medium has 36.