Model comparison
gpt-oss-20b vs Nemotron 3 Ultra
Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 29× less per token, which makes it the better buy when Nemotron 3 Ultra's lead doesn't matter for your workload.
Last verified . 25 shared benchmarks.
NVIDIASummary
- They share 25 benchmarks with published results for both. gpt-oss-20b scores higher in 1 category and Nemotron 3 Ultra in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Nemotron 3 Ultra leads 66.3 to 35.5.
- The biggest single-benchmark swing is GPQA Diamond: 60.8% for gpt-oss-20b and 85.4% for Nemotron 3 Ultra.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $0.60 / $2.40 for Nemotron 3 Ultra.
- Nemotron 3 Ultra accepts more context: 262K tokens versus 131K.
Side by side
| gpt-oss-20b | Nemotron 3 Ultra | |
|---|---|---|
| Provider | OpenAI | NVIDIA |
| Noometry Index | 32.5 | 42.5 |
| Released | 2025-08-05 | 2026-06-04 |
| Weights | Open | Open |
| Context window | 131K | 262K |
| Max output | 16K | 128K |
| Input $ / M tokens | $0.018 | $0.60 |
| Output $ / M tokens | $0.09 | $2.40 |
| Results tracked | 34 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
gpt-oss-20b: 37.6 (#192), Nemotron 3 Ultra: 38.1 (#182)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| SciCode | 34.4% | 40.3% |
| WeirdML | 40.9% | 43.5% |
| LMArena Coding | 1306 | 1468 |
| FrontierCode | — | 13.6% |
| ALE-Bench | 566.05 | — |
Agentic & Tool Use Nemotron 3 Ultra leads
gpt-oss-20b: 9.3 (#154), Nemotron 3 Ultra: 23.4 (#126)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| Terminal-Bench | 3.4% | — |
| APEX-Agents | — | 22.7% |
Reasoning Nemotron 3 Ultra leads
gpt-oss-20b: 19.3 (#261), Nemotron 3 Ultra: 31.0 (#82)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| CritPt | 1.4% | 3.1% |
| Chess Puzzles | 4% | 12% |
| LMArena Hard Prompts | 1274 | 1452 |
| DTBench | 68% | 90.1% |
| LMCA | 14.5% | 36.9% |
| Epoch Capabilities Index | 137.82 | 146.17 |
| Kagi LLM Benchmark | 53.2% | — |
| Mystery Game Puzzles | — | 20% |
Math gpt-oss-20b leads
gpt-oss-20b: 39.4 (#103), Nemotron 3 Ultra: 35.3 (#187)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 65.3% | 86.7% |
| LMArena Math | 1317 | 1457 |
| ProofBench | — | 2% |
| Omni-MATH | 56.5% | — |
Knowledge Nemotron 3 Ultra leads
gpt-oss-20b: 34.6 (#195), Nemotron 3 Ultra: 52.5 (#61)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| GPQA Diamond | 60.8% | 85.4% |
| LMArena Expert | 1258 | 1472 |
| MMLU-Pro | 74% | — |
| GPQA (HELM) | 59.4% | — |
Multilingual Nemotron 3 Ultra leads
gpt-oss-20b: 42.2 (#197), Nemotron 3 Ultra: 53.5 (#64)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| LMArena Non-English | 1268 | 1427 |
| LMArena Chinese | 1314 | 1497 |
| LMArena German | 1255 | 1471 |
| LMArena Korean | 1236 | 1386 |
| LMArena Russian | 1278 | 1417 |
| LMArena Spanish | 1267 | 1454 |
| LMArena French | — | 1472 |
| LMArena Japanese | 1244 | — |
Instruction Following Nemotron 3 Ultra leads
gpt-oss-20b: 61.8 (#240), Nemotron 3 Ultra: 74.7 (#88)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| LMArena Instruction Following | 1236 | 1418 |
| IFEval | 73.2% | — |
Long Context Nemotron 3 Ultra leads
gpt-oss-20b: 37.9 (#209), Nemotron 3 Ultra: 43.9 (#81)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| LMArena Longer Query | 1250 | 1437 |
Writing & Preference Nemotron 3 Ultra leads
gpt-oss-20b: 35.5 (#265), Nemotron 3 Ultra: 66.3 (#36)
| Benchmark | gpt-oss-20b | Nemotron 3 Ultra |
|---|---|---|
| LMArena Text | 1287 | 1445 |
| LMArena Creative Writing | 1201 | 1398 |
| EQ-Bench Creative Writing | 666 | 1692 |
| LMArena Multi-Turn | 1268 | 1411 |
| WildBench | 73.7% | — |
Frequently asked questions
Is gpt-oss-20b better than Nemotron 3 Ultra?
Nemotron 3 Ultra is the stronger model overall, scoring 42.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 29× less per token, which makes it the better buy when Nemotron 3 Ultra's lead doesn't matter for your workload.
Which is cheaper, gpt-oss-20b or Nemotron 3 Ultra?
gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Nemotron 3 Ultra lists at $0.60 and $2.40.
Is gpt-oss-20b or Nemotron 3 Ultra better for coding?
They score almost the same on coding (37.6 vs 38.1); test both on your own repository before choosing.
Which has the bigger context window?
Nemotron 3 Ultra does, with 262K tokens against 131K.
How many benchmarks do gpt-oss-20b and Nemotron 3 Ultra share?
25 benchmarks have published results for both models. gpt-oss-20b has 34 scored results on Noometry and Nemotron 3 Ultra has 30.