# GPT-6 Sol vs gpt-oss-120b

> GPT-6 Sol is the stronger model overall, scoring 49.2 to 36.0 on the Noometry Index. gpt-oss-120b costs 57× less per token, which makes it the better buy when GPT-6 Sol's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-6-sol-vs-gpt-oss-120b
- Last updated: 2026-10-10
- Shared benchmarks: 15

## Summary

- They share 15 benchmarks with published results for both. GPT-6 Sol scores higher in 8 categories and gpt-oss-120b in 1 category; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Sol leads 49.3 to 19.5.
- The biggest single-benchmark swing is Vectara Hallucination Rate: 6.5% for GPT-6 Sol and 14.2% for gpt-oss-120b.
- gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $2 / $10 for GPT-6 Sol.
- GPT-6 Sol accepts more context: 1.05M tokens versus 131K.
- gpt-oss-120b has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 49.2 | 36.0 |
| Rank | 46 | 210 |
| Context | 1.05M | 131K |
| Input $/M | $2 | $0.037 |
| Output $/M | $10 | $0.17 |
| Weights | Proprietary | Open |

## Coding

- GPT-6 Sol: 59.6 (#11)
- gpt-oss-120b: 33.2 (#251)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Coding | 1453 | 1380 |
| DeepSWE | 68.8% | — |
| FrontierCode | 49.3% | — |
| SWE-bench Verified (bash only) | — | 26% |
| Aider Polyglot | — | 41.8% |
| LMArena WebDev | 1692 | — |
| SciCode | — | 36% |
| WeirdML | — | 48.2% |
| ALE-Bench | — | 575.62 |
| AlgoTune | — | 1.41 |

## Agentic & Tool Use

- GPT-6 Sol: 33.5 (#49)
- gpt-oss-120b: 12.1 (#145)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| Terminal-Bench | — | 18.7% |
| APEX-Agents | — | 4.4% |
| GDP.pdf | 28% | — |
| METR Time Horizons | — | 56.6% |
| Vending-Bench 2 | — | -21.53 |

## Reasoning

- GPT-6 Sol: 49.3 (#34)
- gpt-oss-120b: 19.5 (#236)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Hard Prompts | 1420 | 1364 |
| SimpleBench | — | 22.1% |
| Kagi LLM Benchmark | — | 58.6% |
| NYT Connections (extended) | 90.1% | — |
| CritPt | — | 1.1% |
| Chess Puzzles | — | 20% |
| EBR-Bench | 53.3% | — |
| Mystery Game Puzzles | — | 2% |
| DTBench | — | 76.3% |
| LMCA | — | 22.1% |
| Surface Evolver Bench | — | 25% |
| Epoch Capabilities Index | — | 139.94 |

## Math

- GPT-6 Sol: 35.8 (#135)
- gpt-oss-120b: 50.6 (#43)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Math | 1404 | 1388 |
| OTIS Mock AIME 2024-2025 | — | 88.9% |
| Omni-MATH | — | 68.8% |

## Knowledge

- GPT-6 Sol: 43.0 (#87)
- gpt-oss-120b: 42.3 (#89)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| Vectara Hallucination Rate | 6.5% | 14.2% |
| LMArena Expert | 1449 | 1356 |
| GPQA Diamond | — | 75.8% |
| MMLU-Pro | — | 79.5% |
| Confabulations | — | 15.7% |
| GPQA (HELM) | — | 68.4% |

## Multimodal

- GPT-6 Sol: 47.6 (#9)
- gpt-oss-120b: —

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| Furniture Assembly | 58.3% | — |

## Multilingual

- GPT-6 Sol: 51.0 (#113)
- gpt-oss-120b: 48.2 (#142)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Non-English | 1389 | 1350 |
| LMArena Chinese | 1416 | 1384 |
| LMArena Russian | 1407 | 1343 |
| LMArena Spanish | 1377 | 1388 |
| LMArena French | — | 1370 |
| LMArena German | — | 1355 |
| LMArena Japanese | — | 1332 |
| LMArena Korean | — | 1283 |

## Instruction Following

- GPT-6 Sol: 74.4 (#97)
- gpt-oss-120b: 69.5 (#168)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Instruction Following | 1406 | 1318 |
| IFEval | — | 83.6% |

## Long Context

- GPT-6 Sol: 43.2 (#100)
- gpt-oss-120b: 31.4 (#273)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Longer Query | 1415 | 1319 |
| Fiction.LiveBench | — | 44.4% |

## Writing & Preference

- GPT-6 Sol: 72.3 (#15)
- gpt-oss-120b: 46.7 (#213)

| Benchmark | GPT-6 Sol | gpt-oss-120b |
|---|---|---|
| LMArena Text | 1396 | 1366 |
| LMArena Creative Writing | 1392 | 1274 |
| EQ-Bench Creative Writing | 2125 | 961 |
| LMArena Multi-Turn | 1405 | 1340 |
| Short-Story Creative Writing | — | 77.1% |
| WildBench | — | 84.5% |

## FAQ

### Is GPT-6 Sol better than gpt-oss-120b?

GPT-6 Sol is the stronger model overall, scoring 49.2 to 36.0 on the Noometry Index. gpt-oss-120b costs 57× less per token, which makes it the better buy when GPT-6 Sol's lead doesn't matter for your workload.

### Which is cheaper, GPT-6 Sol 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; GPT-6 Sol lists at $2 and $10.

### Is GPT-6 Sol or gpt-oss-120b better for coding?

GPT-6 Sol scores higher on coding benchmarks: 59.6 versus 33.2 in the Noometry coding category.

### Which has the bigger context window?

GPT-6 Sol does, with 1.05M tokens against 131K.

### How many benchmarks do GPT-6 Sol and gpt-oss-120b share?

15 benchmarks have published results for both models. GPT-6 Sol has 22 scored results on Noometry and gpt-oss-120b has 48.