# DeepSeek-V3.1 vs gpt-oss-120b

> DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 36.0 on the Noometry Index. gpt-oss-120b costs 6.0× less per token, which makes it the better buy when DeepSeek-V3.1's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/deepseek-v3-1-vs-gpt-oss-120b
- Last updated: 2026-10-10
- Shared benchmarks: 26

## Summary

- They share 26 benchmarks with published results for both. DeepSeek-V3.1 scores higher in 7 categories and gpt-oss-120b in 1 category; 7 gaps are clear of the uncertainty.
- The widest gap is in math, where gpt-oss-120b leads 50.6 to 36.1.
- The biggest single-benchmark swing is SimpleBench: 40% for DeepSeek-V3.1 and 22.1% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $0.25 / $0.95 for DeepSeek-V3.1.
- DeepSeek-V3.1 accepts more context: 164K tokens versus 131K.

## Snapshot

| | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| Provider | DeepSeek | OpenAI |
| Noometry Index | 42.2 | 36.0 |
| Rank | 106 | 210 |
| Context | 164K | 131K |
| Input $/M | $0.25 | $0.037 |
| Output $/M | $0.95 | $0.17 |
| Weights | Open | Open |

## Coding

- DeepSeek-V3.1: 39.8 (#140)
- gpt-oss-120b: 33.2 (#251)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| WeirdML | 38.4% | 48.2% |
| LMArena Coding | 1417 | 1380 |
| SWE-bench Verified (bash only) | — | 26% |
| Aider Polyglot | — | 41.8% |
| SciCode | — | 36% |
| ALE-Bench | — | 575.62 |
| AlgoTune | — | 1.41 |

## Agentic & Tool Use

- DeepSeek-V3.1: —
- gpt-oss-120b: 12.1 (#145)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| METR Time Horizons | — | 56.6% |
| Vending-Bench 2 | — | -21.53 |

## Reasoning

- DeepSeek-V3.1: 27.3 (#109)
- gpt-oss-120b: 19.5 (#236)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| SimpleBench | 40% | 22.1% |
| Kagi LLM Benchmark | 53.2% | 58.6% |
| LMArena Hard Prompts | 1417 | 1364 |
| DTBench | 82.7% | 76.3% |
| LMCA | 24.3% | 22.1% |
| Epoch Capabilities Index | 139.92 | 139.94 |
| CritPt | — | 1.1% |
| Chess Puzzles | — | 20% |
| Mystery Game Puzzles | — | 2% |
| Surface Evolver Bench | — | 25% |
| ForecastBench | 58 | — |

## Math

- DeepSeek-V3.1: 36.1 (#125)
- gpt-oss-120b: 50.6 (#43)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| LMArena Math | 1420 | 1388 |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| Omni-MATH | — | 68.8% |

## Knowledge

- DeepSeek-V3.1: 43.0 (#86)
- gpt-oss-120b: 42.3 (#89)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| Vectara Hallucination Rate | 5.5% | 14.2% |
| LMArena Expert | 1403 | 1356 |
| GPQA Diamond | — | 75.8% |
| MMLU-Pro | — | 79.5% |
| Confabulations | — | 15.7% |
| GPQA (HELM) | — | 68.4% |

## Multilingual

- DeepSeek-V3.1: 51.8 (#100)
- gpt-oss-120b: 48.2 (#142)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1400 | 1350 |
| LMArena Chinese | 1470 | 1384 |
| LMArena French | 1447 | 1370 |
| LMArena German | 1412 | 1355 |
| LMArena Japanese | 1374 | 1332 |
| LMArena Korean | 1337 | 1283 |
| LMArena Russian | 1405 | 1343 |
| LMArena Spanish | 1429 | 1388 |

## Instruction Following

- DeepSeek-V3.1: 74.1 (#105)
- gpt-oss-120b: 69.5 (#168)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| LMArena Instruction Following | 1400 | 1318 |
| IFEval | — | 83.6% |

## Long Context

- DeepSeek-V3.1: 36.3 (#227)
- gpt-oss-120b: 31.4 (#273)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| Fiction.LiveBench | 52.8% | 44.4% |
| LMArena Longer Query | 1422 | 1319 |

## Writing & Preference

- DeepSeek-V3.1: 60.5 (#93)
- gpt-oss-120b: 46.7 (#213)

| Benchmark | DeepSeek-V3.1 | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1420 | 1366 |
| LMArena Creative Writing | 1401 | 1274 |
| EQ-Bench Creative Writing | 1436 | 961 |
| LMArena Multi-Turn | 1408 | 1340 |
| Short-Story Creative Writing | — | 77.1% |
| WildBench | — | 84.5% |

## FAQ

### Is DeepSeek-V3.1 better than gpt-oss-120b?

DeepSeek-V3.1 is the stronger model overall, scoring 42.2 to 36.0 on the Noometry Index. gpt-oss-120b costs 6.0× less per token, which makes it the better buy when DeepSeek-V3.1's lead doesn't matter for your workload.

### Which is cheaper, DeepSeek-V3.1 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-V3.1 lists at $0.25 and $0.95.

### Is DeepSeek-V3.1 or gpt-oss-120b better for coding?

DeepSeek-V3.1 scores higher on coding benchmarks: 39.8 versus 33.2 in the Noometry coding category.

### Which has the bigger context window?

DeepSeek-V3.1 does, with 164K tokens against 131K.

### How many benchmarks do DeepSeek-V3.1 and gpt-oss-120b share?

26 benchmarks have published results for both models. DeepSeek-V3.1 has 27 scored results on Noometry and gpt-oss-120b has 48.