Model comparison
DeepSeek-R1 vs gpt-oss-120b
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 36.0 on the Noometry Index. gpt-oss-120b costs 13× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Last verified . 39 shared benchmarks.
DeepSeekSummary
- They share 39 benchmarks with published results for both. DeepSeek-R1 scores higher in 7 categories and gpt-oss-120b in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where DeepSeek-R1 leads 30.4 to 12.1.
- The biggest single-benchmark swing is Fiction.LiveBench: 75% for DeepSeek-R1 and 44.4% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $0.50 / $2.15 for DeepSeek-R1.
- DeepSeek-R1 accepts more context: 164K tokens versus 131K.
- gpt-oss-120b has downloadable open weights; the other is API-only.
Side by side
| DeepSeek-R1 | gpt-oss-120b | |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.0 | 36.0 |
| Released | 2025-01-20 | 2025-08-05 |
| Weights | Proprietary | Open |
| Context window | 164K | 131K |
| Max output | 64K | 41K |
| Input $ / M tokens | $0.50 | $0.037 |
| Output $ / M tokens | $2.15 | $0.17 |
| Results tracked | 52 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek-R1 leads
DeepSeek-R1: 45.9 (#63), gpt-oss-120b: 33.2 (#251)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| Aider Polyglot | 71.4% | 41.8% |
| SciCode | 35.7% | 36% |
| WeirdML | 41.6% | 48.2% |
| LMArena Coding | 1427 | 1380 |
| ALE-Bench | 804.12 | 575.62 |
| AlgoTune | 1.7 | 1.41 |
| SWE-bench Verified (bash only) | — | 26% |
| LiveBench Coding | 66.7% | — |
Agentic & Tool Use DeepSeek-R1 leads
DeepSeek-R1: 30.4 (#68), gpt-oss-120b: 12.1 (#145)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| METR Time Horizons | 53.8% | 56.6% |
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| DeepResearch Bench | 35.1% | — |
| BALROG | 34.9% | — |
| Vending-Bench 2 | — | -21.53 |
Reasoning Too close to call
DeepSeek-R1: 18.7 (#253), gpt-oss-120b: 19.5 (#236)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| SimpleBench | 40.8% | 22.1% |
| Kagi LLM Benchmark | 69.4% | 58.6% |
| CritPt | 1.1% | 1.1% |
| LMArena Hard Prompts | 1415 | 1364 |
| Epoch Capabilities Index | 141.3 | 139.94 |
| ARC-AGI-2 | 1.3% | — |
| ARC-AGI-1 | 21.2% | — |
| Chess Puzzles | — | 20% |
| LiveBench Reasoning | 83.2% | — |
| Mystery Game Puzzles | — | 2% |
| DTBench | — | 76.3% |
| LiveBench Data Analysis | 69.8% | — |
| LMCA | — | 22.1% |
| Surface Evolver Bench | — | 25% |
| ForecastBench | 60 | — |
| LiveBench | 71.6% | — |
Math gpt-oss-120b leads
DeepSeek-R1: 42.0 (#70), gpt-oss-120b: 50.6 (#43)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 66.4% | 88.9% |
| Omni-MATH | 42.4% | 68.8% |
| LMArena Math | 1401 | 1388 |
| LiveBench Math | 80.7% | — |
| MATH Level 5 | 96.6% | — |
Knowledge DeepSeek-R1 leads
DeepSeek-R1: 44.3 (#81), gpt-oss-120b: 42.3 (#89)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| GPQA Diamond | 76.3% | 75.8% |
| MMLU-Pro | 79.3% | 79.5% |
| Confabulations | 12.7% | 15.7% |
| Vectara Hallucination Rate | 11.3% | 14.2% |
| GPQA (HELM) | 66.6% | 68.4% |
| LMArena Expert | 1394 | 1356 |
Multilingual DeepSeek-R1 leads
DeepSeek-R1: 52.6 (#80), gpt-oss-120b: 48.2 (#142)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1412 | 1350 |
| LMArena Chinese | 1443 | 1384 |
| LMArena French | 1415 | 1370 |
| LMArena German | 1405 | 1355 |
| LMArena Japanese | 1394 | 1332 |
| LMArena Korean | 1359 | 1283 |
| LMArena Russian | 1423 | 1343 |
| LMArena Spanish | 1410 | 1388 |
Instruction Following DeepSeek-R1 leads
DeepSeek-R1: 72.2 (#138), gpt-oss-120b: 69.5 (#168)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| IFEval | 78.4% | 83.6% |
| LMArena Instruction Following | 1382 | 1318 |
| LiveBench Instruction Following | 80.5% | — |
Long Context DeepSeek-R1 leads
DeepSeek-R1: 45.4 (#36), gpt-oss-120b: 31.4 (#273)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | 75% | 44.4% |
| LMArena Longer Query | 1391 | 1319 |
Writing & Preference DeepSeek-R1 leads
DeepSeek-R1: 61.5 (#84), gpt-oss-120b: 46.7 (#213)
| Benchmark | DeepSeek-R1 | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1428 | 1366 |
| LMArena Creative Writing | 1406 | 1274 |
| Short-Story Creative Writing | 83% | 77.1% |
| EQ-Bench Creative Writing | 1500 | 961 |
| WildBench | 82.8% | 84.5% |
| LMArena Multi-Turn | 1404 | 1340 |
| LiveBench Language | 48.5% | — |
Frequently asked questions
Is DeepSeek-R1 better than gpt-oss-120b?
DeepSeek-R1 is the stronger model overall, scoring 42.0 to 36.0 on the Noometry Index. gpt-oss-120b costs 13× less per token, which makes it the better buy when DeepSeek-R1's lead doesn't matter for your workload.
Which is cheaper, DeepSeek-R1 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; DeepSeek-R1 lists at $0.50 and $2.15.
Is DeepSeek-R1 or gpt-oss-120b better for coding?
DeepSeek-R1 scores higher on coding benchmarks: 45.9 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
DeepSeek-R1 does, with 164K tokens against 131K.
How many benchmarks do DeepSeek-R1 and gpt-oss-120b share?
39 benchmarks have published results for both models. DeepSeek-R1 has 52 scored results on Noometry and gpt-oss-120b has 48.