Model comparison
GPT-5 Nano vs gpt-oss-120b
gpt-oss-120b is the stronger model overall, scoring 36.3 to 33.5 on the Noometry Index.
Last verified . 36 shared benchmarks.
Summary
- They share 36 benchmarks with published results for both. GPT-5 Nano scores higher in 3 categories and gpt-oss-120b in 6 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 52.5 to 29.4.
- The biggest single-benchmark swing is LMCA: 7.9% for GPT-5 Nano and 22.1% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $0.05 / $0.40 for GPT-5 Nano.
- GPT-5 Nano accepts more context: 400K tokens versus 131K.
- gpt-oss-120b has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | gpt-oss-120b | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 33.5 | 36.3 |
| Released | 2025-08-07 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 41K |
| Input $ / M tokens | $0.05 | $0.037 |
| Output $ / M tokens | $0.40 | $0.17 |
| Results tracked | 49 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
GPT-5 Nano: 33.6 (#254), gpt-oss-120b: 33.5 (#256)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| SWE-bench Verified (bash only) | 34.8% | 26% |
| WeirdML | 38.1% | 48.2% |
| LMArena Coding | 1351 | 1380 |
| ALE-Bench | 718.67 | 575.62 |
| Aider Polyglot | — | 41.8% |
| SciCode | — | 36% |
| AlgoTune | — | 1.41 |
Agentic & Tool Use GPT-5 Nano leads
GPT-5 Nano: 25.8 (#106), gpt-oss-120b: 12.2 (#153)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | 21.8% | 18.7% |
| APEX-Agents | — | 4.4% |
| Berkeley Function Calling Leaderboard | 51.5% | — |
| METR Time Horizons | — | 56.6% |
| Vending-Bench 2 | — | -21.53 |
Reasoning gpt-oss-120b leads
GPT-5 Nano: 16.3 (#306), gpt-oss-120b: 20.0 (#245)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| Kagi LLM Benchmark | 62.2% | 58.6% |
| Chess Puzzles | 27% | 20% |
| LMArena Hard Prompts | 1328 | 1364 |
| Mystery Game Puzzles | 9% | 2% |
| DTBench | 62.7% | 76.3% |
| LMCA | 7.9% | 22.1% |
| Epoch Capabilities Index | 139.38 | 139.93 |
| ARC-AGI-2 | 2.6% | — |
| SimpleBench | — | 22.1% |
| ARC-AGI-1 | 20.7% | — |
| CritPt | — | 1.1% |
| Surface Evolver Bench | — | 25% |
| ForecastBench | 59.1 | — |
Math gpt-oss-120b leads
GPT-5 Nano: 29.4 (#241), gpt-oss-120b: 52.5 (#50)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.1% | 88.9% |
| Omni-MATH | 54.6% | 68.8% |
| LMArena Math | 1317 | 1389 |
| FrontierMath (Tiers 1-3) | 20% | — |
| FrontierMath Tier 4 | 2.4% | — |
| ProofBench | 12% | — |
| MATH Level 5 | 95.2% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge gpt-oss-120b leads
GPT-5 Nano: 35.9 (#178), gpt-oss-120b: 42.4 (#96)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 69.4% | 75.8% |
| MMLU-Pro | 77.8% | 79.5% |
| Vectara Hallucination Rate | 10.5% | 14.2% |
| GPQA (HELM) | 67.9% | 68.4% |
| LMArena Expert | 1321 | 1356 |
| SimpleQA Verified | 11.7% | — |
| Confabulations | — | 15.7% |
Multimodal Not comparable
GPT-5 Nano: 31.3 (#108), gpt-oss-120b: —
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| LMArena Vision | 1159 | — |
| VPCT | 37.2% | — |
Multilingual gpt-oss-120b leads
GPT-5 Nano: 45.3 (#172), gpt-oss-120b: 48.0 (#147)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1313 | 1351 |
| LMArena Chinese | 1356 | 1385 |
| LMArena German | 1327 | 1353 |
| LMArena Japanese | 1226 | 1331 |
| LMArena Korean | 1269 | 1282 |
| LMArena Russian | 1296 | 1343 |
| LMArena Spanish | 1360 | 1389 |
| LMArena French | — | 1369 |
Instruction Following GPT-5 Nano leads
GPT-5 Nano: 75.0 (#79), gpt-oss-120b: 69.3 (#173)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| IFEval | 93.2% | 83.6% |
| LMArena Instruction Following | 1306 | 1318 |
Long Context Too close to call
GPT-5 Nano: 31.3 (#281), gpt-oss-120b: 31.4 (#278)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | 44.4% | 44.4% |
| LMArena Longer Query | 1312 | 1319 |
Writing & Preference gpt-oss-120b leads
GPT-5 Nano: 39.1 (#249), gpt-oss-120b: 46.5 (#217)
| Benchmark | GPT-5 Nano | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1320 | 1365 |
| LMArena Creative Writing | 1249 | 1275 |
| EQ-Bench Creative Writing | 705 | 961 |
| WildBench | 80.6% | 84.5% |
| LMArena Multi-Turn | 1311 | 1340 |
| Short-Story Creative Writing | — | 77.1% |
Frequently asked questions
Is GPT-5 Nano better than gpt-oss-120b?
gpt-oss-120b is the stronger model overall, scoring 36.3 to 33.5 on the Noometry Index.
Which is cheaper, GPT-5 Nano 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 Nano lists at $0.05 and $0.40.
Is GPT-5 Nano or gpt-oss-120b better for coding?
They score almost the same on coding (33.6 vs 33.5); test both on your own repository before choosing.
Which has the bigger context window?
GPT-5 Nano does, with 400K tokens against 131K.
How many benchmarks do GPT-5 Nano and gpt-oss-120b share?
36 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and gpt-oss-120b has 48.