Model comparison
gpt-oss-20b vs Mistral Medium
Mistral Medium is the stronger model overall, scoring 36.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 83× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Last verified . 25 shared benchmarks.
Mistral AISummary
- They share 25 benchmarks with published results for both. gpt-oss-20b scores higher in 3 categories and Mistral Medium in 6 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Mistral Medium leads 60.0 to 35.5.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 65.3% for gpt-oss-20b and 32.2% for Mistral Medium.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $1.50 / $7.50 for Mistral Medium.
- Mistral Medium accepts more context: 262K tokens versus 131K.
Side by side
| gpt-oss-20b | Mistral Medium | |
|---|---|---|
| Provider | OpenAI | Mistral AI |
| Noometry Index | 32.5 | 36.3 |
| Released | 2025-08-05 | 2023-12-11 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 16K | 262K |
| Input $ / M tokens | $0.018 | $1.50 |
| Output $ / M tokens | $0.09 | $7.50 |
| Results tracked | 34 | 36 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-20b leads
gpt-oss-20b: 37.6 (#192), Mistral Medium: 34.2 (#243)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| SciCode | 34.4% | 40.2% |
| WeirdML | 40.9% | 43.7% |
| LMArena Coding | 1306 | 1434 |
| ALE-Bench | 566.05 | 763.98 |
| FrontierCode | — | 8% |
Agentic & Tool Use Mistral Medium leads
gpt-oss-20b: 9.3 (#154), Mistral Medium: 28.3 (#90)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| Terminal-Bench | 3.4% | — |
| Berkeley Function Calling Leaderboard | — | 37.7% |
Reasoning Mistral Medium leads
gpt-oss-20b: 19.3 (#261), Mistral Medium: 24.0 (#167)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| Kagi LLM Benchmark | 53.2% | 50% |
| CritPt | 1.4% | 0% |
| LMArena Hard Prompts | 1274 | 1426 |
| DTBench | 68% | 75.5% |
| LMCA | 14.5% | 26.1% |
| Chess Puzzles | 4% | — |
| Surface Evolver Bench | — | 26.9% |
| Epoch Capabilities Index | 137.82 | — |
Math gpt-oss-20b leads
gpt-oss-20b: 39.4 (#103), Mistral Medium: 28.1 (#245)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | 32.2% |
| LMArena Math | 1317 | 1408 |
| ProofBench | — | 9% |
| Omni-MATH | 56.5% | — |
| MATH Level 5 | — | 81.6% |
| FrontierMath (Feb 2025 set) | — | 0.3% |
Knowledge gpt-oss-20b leads
gpt-oss-20b: 34.6 (#195), Mistral Medium: 25.0 (#265)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| GPQA Diamond | 60.8% | 59.5% |
| LMArena Expert | 1258 | 1408 |
| Humanity's Last Exam | — | 4.5% |
| MMLU-Pro | 74% | — |
| Vectara Hallucination Rate | — | 22.7% |
| GPQA (HELM) | 59.4% | — |
Multimodal Not comparable
gpt-oss-20b: —, Mistral Medium: 35.3 (#88)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| LMArena Vision | — | 1172 |
Multilingual Mistral Medium leads
gpt-oss-20b: 42.2 (#197), Mistral Medium: 52.1 (#91)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| LMArena Non-English | 1268 | 1408 |
| LMArena Chinese | 1314 | 1447 |
| LMArena German | 1255 | 1432 |
| LMArena Japanese | 1244 | 1378 |
| LMArena Korean | 1236 | 1380 |
| LMArena Russian | 1278 | 1411 |
| LMArena Spanish | 1267 | 1433 |
| LMArena French | — | 1459 |
Instruction Following Mistral Medium leads
gpt-oss-20b: 61.8 (#240), Mistral Medium: 73.7 (#116)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| LMArena Instruction Following | 1236 | 1398 |
| IFEval | 73.2% | — |
Long Context Mistral Medium leads
gpt-oss-20b: 37.9 (#209), Mistral Medium: 42.9 (#114)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| LMArena Longer Query | 1250 | 1406 |
Writing & Preference Mistral Medium leads
gpt-oss-20b: 35.5 (#265), Mistral Medium: 60.0 (#103)
| Benchmark | gpt-oss-20b | Mistral Medium |
|---|---|---|
| LMArena Text | 1287 | 1424 |
| LMArena Creative Writing | 1201 | 1391 |
| LMArena Multi-Turn | 1268 | 1418 |
| Short-Story Creative Writing | — | 77.3% |
| EQ-Bench Creative Writing | 666 | — |
| WildBench | 73.7% | — |
Frequently asked questions
Is gpt-oss-20b better than Mistral Medium?
Mistral Medium is the stronger model overall, scoring 36.3 to 32.5 on the Noometry Index. gpt-oss-20b costs 83× less per token, which makes it the better buy when Mistral Medium's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or Mistral Medium?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Mistral Medium lists at $1.50 and $7.50.
Is gpt-oss-20b or Mistral Medium better for coding?
gpt-oss-20b scores higher on coding benchmarks: 37.6 versus 34.2 in the Noometry coding category.
Which has the bigger context window?
Mistral Medium does, with 262K tokens against 131K.
How many benchmarks do gpt-oss-20b and Mistral Medium share?
25 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Mistral Medium has 36.