Model comparison
Claude Sonnet 4 vs gpt-oss-120b
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 85× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Last verified . 41 shared benchmarks.
Summary
- They share 41 benchmarks with published results for both. Claude Sonnet 4 scores higher in 6 categories and gpt-oss-120b in 3 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where Claude Sonnet 4 leads 38.5 to 12.2.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 64.9% for Claude Sonnet 4 and 26% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.
- Claude Sonnet 4 accepts more context: 200K tokens versus 131K.
- gpt-oss-120b has downloadable open weights; the other is API-only.
Side by side
| Claude Sonnet 4 | gpt-oss-120b | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 40.8 | 36.3 |
| Released | 2025-05-22 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 200K | 131K |
| Max output | 64K | 41K |
| Input $ / M tokens | $3 | $0.037 |
| Output $ / M tokens | $15 | $0.17 |
| Results tracked | 58 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Sonnet 4 leads
Claude Sonnet 4: 43.5 (#88), gpt-oss-120b: 33.5 (#256)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| SWE-bench Verified (bash only) | 64.9% | 26% |
| Aider Polyglot | 61.3% | 41.8% |
| SciCode | 40% | 36% |
| WeirdML | 46.1% | 48.2% |
| LMArena Coding | 1414 | 1380 |
| ALE-Bench | 655.35 | 575.62 |
| GSO | 4.9% | — |
| AlgoTune | — | 1.41 |
Agentic & Tool Use Claude Sonnet 4 leads
Claude Sonnet 4: 38.5 (#31), gpt-oss-120b: 12.2 (#153)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| METR Time Horizons | 62% | 56.6% |
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| TheAgentCompany | 33.1% | — |
| Cybench | 35% | — |
| DeepResearch Bench | 46.6% | — |
| OSWorld | 43.9% | — |
| Vending-Bench 2 | — | -21.53 |
Reasoning Claude Sonnet 4 leads
Claude Sonnet 4: 22.9 (#187), gpt-oss-120b: 20.0 (#245)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| SimpleBench | 45.5% | 22.1% |
| Kagi LLM Benchmark | 73% | 58.6% |
| CritPt | 0.3% | 1.1% |
| LMArena Hard Prompts | 1372 | 1364 |
| DTBench | 77.1% | 76.3% |
| LMCA | 29% | 22.1% |
| Epoch Capabilities Index | 141.69 | 139.93 |
| ARC-AGI-2 | 5.9% | — |
| ARC-AGI-1 | 40% | — |
| Chess Puzzles | — | 20% |
| EnigmaEval | 3.1% | — |
| Mystery Game Puzzles | — | 2% |
| Surface Evolver Bench | — | 25% |
| ForecastBench | 60.2 | — |
Math gpt-oss-120b leads
Claude Sonnet 4: 43.3 (#80), gpt-oss-120b: 52.5 (#50)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 71.1% | 88.9% |
| Omni-MATH | 60.2% | 68.8% |
| LMArena Math | 1375 | 1389 |
| MATH Level 5 | 84.4% | — |
| FrontierMath (Feb 2025 set) | 4.1% | — |
| FrontierMath Tier 4 (v1) | 0% | — |
Knowledge Too close to call
Claude Sonnet 4: 41.8 (#108), gpt-oss-120b: 42.4 (#96)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 79.2% | 75.8% |
| MMLU-Pro | 84.3% | 79.5% |
| Confabulations | 13.2% | 15.7% |
| Vectara Hallucination Rate | 10.3% | 14.2% |
| GPQA (HELM) | 70.6% | 68.4% |
| LMArena Expert | 1372 | 1356 |
| Humanity's Last Exam | 7.8% | — |
Multimodal Not comparable
Claude Sonnet 4: 26.2 (#121), gpt-oss-120b: —
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| LMArena Vision | 1191 | — |
| GeoBench | 37% | — |
| VPCT | 34% | — |
| MindCube | 44.8% | — |
Multilingual gpt-oss-120b leads
Claude Sonnet 4: 46.7 (#156), gpt-oss-120b: 48.0 (#147)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1333 | 1351 |
| LMArena Chinese | 1350 | 1385 |
| LMArena French | 1363 | 1369 |
| LMArena German | 1331 | 1353 |
| LMArena Japanese | 1302 | 1331 |
| LMArena Korean | 1291 | 1282 |
| LMArena Russian | 1355 | 1343 |
| LMArena Spanish | 1357 | 1389 |
Instruction Following Claude Sonnet 4 leads
Claude Sonnet 4: 71.7 (#145), gpt-oss-120b: 69.3 (#173)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| IFEval | 84% | 83.6% |
| LMArena Instruction Following | 1376 | 1318 |
Long Context Claude Sonnet 4 leads
Claude Sonnet 4: 33.7 (#259), gpt-oss-120b: 31.4 (#278)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | 46.9% | 44.4% |
| LMArena Longer Query | 1398 | 1319 |
Writing & Preference Claude Sonnet 4 leads
Claude Sonnet 4: 57.1 (#132), gpt-oss-120b: 46.5 (#217)
| Benchmark | Claude Sonnet 4 | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1351 | 1365 |
| LMArena Creative Writing | 1345 | 1275 |
| Short-Story Creative Writing | 81.4% | 77.1% |
| EQ-Bench Creative Writing | 1483 | 961 |
| WildBench | 83.8% | 84.5% |
| LMArena Multi-Turn | 1376 | 1340 |
Frequently asked questions
Is Claude Sonnet 4 better than gpt-oss-120b?
Claude Sonnet 4 is the stronger model overall, scoring 40.8 to 36.3 on the Noometry Index. gpt-oss-120b costs 85× less per token, which makes it the better buy when Claude Sonnet 4's lead doesn't matter for your workload.
Which is cheaper, Claude Sonnet 4 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; Claude Sonnet 4 lists at $3 and $15.
Is Claude Sonnet 4 or gpt-oss-120b better for coding?
Claude Sonnet 4 scores higher on coding benchmarks: 43.5 versus 33.5 in the Noometry coding category.
Which has the bigger context window?
Claude Sonnet 4 does, with 200K tokens against 131K.
How many benchmarks do Claude Sonnet 4 and gpt-oss-120b share?
41 benchmarks have published results for both models. Claude Sonnet 4 has 58 scored results on Noometry and gpt-oss-120b has 48.