Model comparison
GPT-5 Nano vs gpt-oss-20b
GPT-5 Nano is the stronger model overall, scoring 33.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 3.8× less per token, which makes it the better buy when GPT-5 Nano's lead doesn't matter for your workload.
Last verified . 32 shared benchmarks.
Summary
- They share 32 benchmarks with published results for both. GPT-5 Nano scores higher in 5 categories and gpt-oss-20b in 4 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5 Nano leads 25.8 to 9.3.
- The biggest single-benchmark swing is Chess Puzzles: 27% for GPT-5 Nano and 4% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 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-20b has downloadable open weights; the other is API-only.
Side by side
| GPT-5 Nano | gpt-oss-20b | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 33.5 | 32.5 |
| Released | 2025-08-07 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 400K | 131K |
| Max output | 128K | 16K |
| Input $ / M tokens | $0.05 | $0.018 |
| Output $ / M tokens | $0.40 | $0.09 |
| Results tracked | 49 | 34 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding gpt-oss-20b leads
GPT-5 Nano: 33.6 (#254), gpt-oss-20b: 37.6 (#192)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| WeirdML | 38.1% | 40.9% |
| LMArena Coding | 1351 | 1306 |
| ALE-Bench | 718.67 | 566.05 |
| SWE-bench Verified (bash only) | 34.8% | — |
| SciCode | — | 34.4% |
Agentic & Tool Use GPT-5 Nano leads
GPT-5 Nano: 25.8 (#106), gpt-oss-20b: 9.3 (#154)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | 21.8% | 3.4% |
| Berkeley Function Calling Leaderboard | 51.5% | — |
Reasoning gpt-oss-20b leads
GPT-5 Nano: 16.3 (#306), gpt-oss-20b: 19.3 (#261)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 62.2% | 53.2% |
| Chess Puzzles | 27% | 4% |
| LMArena Hard Prompts | 1328 | 1274 |
| DTBench | 62.7% | 68% |
| LMCA | 7.9% | 14.5% |
| Epoch Capabilities Index | 139.38 | 137.82 |
| ARC-AGI-2 | 2.6% | — |
| ARC-AGI-1 | 20.7% | — |
| CritPt | — | 1.4% |
| Mystery Game Puzzles | 9% | — |
| ForecastBench | 59.1 | — |
Math gpt-oss-20b leads
GPT-5 Nano: 29.4 (#241), gpt-oss-20b: 39.4 (#103)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 81.1% | 65.3% |
| Omni-MATH | 54.6% | 56.5% |
| LMArena Math | 1317 | 1317 |
| 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-5 Nano leads
GPT-5 Nano: 35.9 (#178), gpt-oss-20b: 34.6 (#195)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| GPQA Diamond | 69.4% | 60.8% |
| MMLU-Pro | 77.8% | 74% |
| GPQA (HELM) | 67.9% | 59.4% |
| LMArena Expert | 1321 | 1258 |
| SimpleQA Verified | 11.7% | — |
| Vectara Hallucination Rate | 10.5% | — |
Multimodal Not comparable
GPT-5 Nano: 31.3 (#108), gpt-oss-20b: —
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| LMArena Vision | 1159 | — |
| VPCT | 37.2% | — |
Multilingual GPT-5 Nano leads
GPT-5 Nano: 45.3 (#172), gpt-oss-20b: 42.2 (#197)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1313 | 1268 |
| LMArena Chinese | 1356 | 1314 |
| LMArena German | 1327 | 1255 |
| LMArena Japanese | 1226 | 1244 |
| LMArena Korean | 1269 | 1236 |
| LMArena Russian | 1296 | 1278 |
| LMArena Spanish | 1360 | 1267 |
Instruction Following GPT-5 Nano leads
GPT-5 Nano: 75.0 (#79), gpt-oss-20b: 61.8 (#240)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| IFEval | 93.2% | 73.2% |
| LMArena Instruction Following | 1306 | 1236 |
Long Context gpt-oss-20b leads
GPT-5 Nano: 31.3 (#281), gpt-oss-20b: 37.9 (#209)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1312 | 1250 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference GPT-5 Nano leads
GPT-5 Nano: 39.1 (#249), gpt-oss-20b: 35.5 (#265)
| Benchmark | GPT-5 Nano | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1320 | 1287 |
| LMArena Creative Writing | 1249 | 1201 |
| EQ-Bench Creative Writing | 705 | 666 |
| WildBench | 80.6% | 73.7% |
| LMArena Multi-Turn | 1311 | 1268 |
Frequently asked questions
Is GPT-5 Nano better than gpt-oss-20b?
GPT-5 Nano is the stronger model overall, scoring 33.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 3.8× less per token, which makes it the better buy when GPT-5 Nano's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or gpt-oss-20b?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; GPT-5 Nano lists at $0.05 and $0.40.
Is GPT-5 Nano or gpt-oss-20b better for coding?
gpt-oss-20b scores higher on coding benchmarks: 37.6 versus 33.6 in the Noometry coding category.
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-20b share?
32 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and gpt-oss-20b has 34.