Model comparison
GPT-5 Mini vs gpt-oss-120b
GPT-5 Mini is the stronger model overall, scoring 41.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 9.8× less per token, which makes it the better buy when GPT-5 Mini's lead doesn't matter for your workload.
Last verified . 43 shared benchmarks.
Summary
- They share 43 benchmarks with published results for both. GPT-5 Mini scores higher in 8 categories and gpt-oss-120b in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5 Mini leads 31.1 to 12.2.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 59.8% for GPT-5 Mini and 26% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $0.25 / $2 for GPT-5 Mini.
- GPT-5 Mini accepts more context: 400K tokens versus 131K.
- gpt-oss-120b has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Mini | gpt-oss-120b | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 41.8 | 36.3 |
| Released | 2025-08-07 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 41K |
| Input $ / M tokens | $0.25 | $0.037 |
| Output $ / M tokens | $2 | $0.17 |
| Results tracked | 60 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5 Mini leads
GPT-5 Mini: 40.1 (#146), gpt-oss-120b: 33.5 (#256)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| SWE-bench Verified (bash only) | 59.8% | 26% |
| SciCode | 39.2% | 36% |
| WeirdML | 52.7% | 48.2% |
| LMArena Coding | 1406 | 1380 |
| ALE-Bench | 799.77 | 575.62 |
| AlgoTune | 1.38 | 1.41 |
| SWE-bench Verified | 64.7% | — |
| Aider Polyglot | — | 41.8% |
| SWE-bench Multilingual | 39.7% | — |
Agentic & Tool Use GPT-5 Mini leads
GPT-5 Mini: 31.1 (#70), gpt-oss-120b: 12.2 (#153)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | 34.8% | 18.7% |
| Vending-Bench 2 | -31.18 | -21.53 |
| APEX-Agents | — | 4.4% |
| Berkeley Function Calling Leaderboard | 55.5% | — |
| METR Time Horizons | — | 56.6% |
Reasoning GPT-5 Mini leads
GPT-5 Mini: 23.9 (#168), gpt-oss-120b: 20.0 (#245)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| Kagi LLM Benchmark | 70.3% | 58.6% |
| CritPt | 0% | 1.1% |
| Chess Puzzles | 30% | 20% |
| LMArena Hard Prompts | 1380 | 1364 |
| Mystery Game Puzzles | 10% | 2% |
| DTBench | 80.5% | 76.3% |
| LMCA | 34.2% | 22.1% |
| Epoch Capabilities Index | 145.52 | 139.93 |
| ARC-AGI-2 | 4.4% | — |
| SimpleBench | — | 22.1% |
| ARC-AGI-1 | 54.3% | — |
| EnigmaEval | 8.2% | — |
| Surface Evolver Bench | — | 25% |
| ForecastBench | 61 | — |
Math gpt-oss-120b leads
GPT-5 Mini: 46.7 (#69), gpt-oss-120b: 52.5 (#50)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 86.7% | 88.9% |
| Omni-MATH | 72.2% | 68.8% |
| LMArena Math | 1378 | 1389 |
| FrontierMath (Tiers 1-3) | 46.7% | — |
| FrontierMath Tier 4 | 12.2% | — |
| ProofBench | 9% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 27.2% | — |
| FrontierMath Tier 4 (v1) | 6.3% | — |
Knowledge GPT-5 Mini leads
GPT-5 Mini: 45.6 (#86), gpt-oss-120b: 42.4 (#96)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 75% | 75.8% |
| MMLU-Pro | 83.5% | 79.5% |
| Confabulations | 13.3% | 15.7% |
| Vectara Hallucination Rate | 12.9% | 14.2% |
| GPQA (HELM) | 75.6% | 68.4% |
| LMArena Expert | 1379 | 1356 |
| Humanity's Last Exam | 19.4% | — |
| SimpleQA Verified | 21.6% | — |
Multimodal Not comparable
GPT-5 Mini: 35.6 (#85), gpt-oss-120b: —
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| LMArena Vision | 1202 | — |
| VPCT | 40.2% | — |
Multilingual Too close to call
GPT-5 Mini: 48.9 (#137), gpt-oss-120b: 48.0 (#147)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1363 | 1351 |
| LMArena Chinese | 1385 | 1385 |
| LMArena French | 1386 | 1369 |
| LMArena German | 1366 | 1353 |
| LMArena Japanese | 1341 | 1331 |
| LMArena Korean | 1308 | 1282 |
| LMArena Russian | 1362 | 1343 |
| LMArena Spanish | 1355 | 1389 |
Instruction Following GPT-5 Mini leads
GPT-5 Mini: 76.2 (#46), gpt-oss-120b: 69.3 (#173)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| IFEval | 92.7% | 83.6% |
| LMArena Instruction Following | 1357 | 1318 |
Long Context GPT-5 Mini leads
GPT-5 Mini: 41.9 (#132), gpt-oss-120b: 31.4 (#278)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | 69.4% | 44.4% |
| LMArena Longer Query | 1355 | 1319 |
Writing & Preference GPT-5 Mini leads
GPT-5 Mini: 55.2 (#148), gpt-oss-120b: 46.5 (#217)
| Benchmark | GPT-5 Mini | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1373 | 1365 |
| LMArena Creative Writing | 1325 | 1275 |
| Short-Story Creative Writing | 83.1% | 77.1% |
| EQ-Bench Creative Writing | 1313 | 961 |
| WildBench | 85.5% | 84.5% |
| LMArena Multi-Turn | 1363 | 1340 |
Frequently asked questions
Is GPT-5 Mini better than gpt-oss-120b?
GPT-5 Mini is the stronger model overall, scoring 41.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 9.8× less per token, which makes it the better buy when GPT-5 Mini's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Mini or gpt-oss-120b?
gpt-oss-120b is cheaper. It lists at $0.037 per million input tokens and $0.17 per million output tokens; GPT-5 Mini lists at $0.25 and $2.
Is GPT-5 Mini or gpt-oss-120b better for coding?
GPT-5 Mini scores higher on coding benchmarks: 40.1 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5 Mini does, with 400K tokens against 131K.
How many benchmarks do GPT-5 Mini and gpt-oss-120b share?
43 benchmarks have published results for both models. GPT-5 Mini has 60 scored results on Noometry and gpt-oss-120b has 48.