# GPT-6 Sol vs Qwen3.8 27B

> GPT-6 Sol is the stronger model overall, scoring 61.8 to 46.0 on the Noometry Index. Qwen3.8 27B costs 6.6× 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-qwen3-8-27b
- Last updated: 2026-10-12
- Shared benchmarks: 30

## Summary

- They share 30 benchmarks with published results for both. GPT-6 Sol scores higher in 7 categories and Qwen3.8 27B in 3 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-6 Sol leads 87.2 to 37.1.
- The biggest single-benchmark swing is ProofBench: 83% for GPT-6 Sol and 16% for Qwen3.8 27B.
- Qwen3.8 27B is cheaper at $0.04 / $2.30 per million input/output tokens, against $2 / $10 for GPT-6 Sol.
- GPT-6 Sol accepts more context: 1.05M tokens versus 262K.
- Qwen3.8 27B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 61.8 | 46.0 |
| Rank | 12 | 68 |
| Context | 1.05M | 262K |
| Input $/M | $2 | $0.04 |
| Output $/M | $10 | $2.30 |
| Weights | Proprietary | Open |

## Coding

- GPT-6 Sol: 60.1 (#11)
- Qwen3.8 27B: 50.5 (#44)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena WebDev | 1688 | 1593 |
| SciCode | 57.6% | 46.6% |
| LMArena Coding | 1447 | 1482 |
| DeepSWE | 68.8% | — |
| FrontierCode | 49.3% | — |
| ALE-Bench | 2,462 | — |

## Agentic & Tool Use

- GPT-6 Sol: 37.2 (#36)
- Qwen3.8 27B: 32.9 (#57)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| APEX-Agents | 54.3% | 47.5% |
| GDP.pdf | 26.4% | — |
| Vending-Bench 2 | 14,428 | — |

## Reasoning

- GPT-6 Sol: 74.0 (#9)
- Qwen3.8 27B: 41.0 (#54)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| ARC-AGI-2 | 89.6% | 42.4% |
| NYT Connections (extended) | 90.1% | 54.5% |
| ARC-AGI-1 | 95.5% | 87.5% |
| CritPt | 30.9% | 5.4% |
| LMArena Hard Prompts | 1418 | 1460 |
| DTBench | 97.3% | 88% |
| LMCA | 59.1% | 41.4% |
| Epoch Capabilities Index | 162.72 | 149.38 |
| EBR-Bench | 53.3% | — |
| Mystery Game Puzzles | 56% | — |
| Surface Evolver Bench | — | 45% |

## Math

- GPT-6 Sol: 87.2 (#7)
- Qwen3.8 27B: 37.1 (#161)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| ProofBench | 83% | 16% |
| LMArena Math | 1402 | 1456 |
| FrontierMath (Tiers 1-3) | 89.8% | — |
| FrontierMath Tier 4 | 90% | — |
| OTIS Mock AIME 2024-2025 | 100% | — |

## Knowledge

- GPT-6 Sol: 64.8 (#15)
- Qwen3.8 27B: 41.6 (#109)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Expert | 1439 | 1482 |
| GPQA Diamond | 94.3% | — |
| SimpleQA Verified | 60.7% | — |
| Vectara Hallucination Rate | 6.5% | — |

## Multimodal

- GPT-6 Sol: 47.6 (#10)
- Qwen3.8 27B: 41.3 (#37)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Vision | 1245 | 1271 |
| Blueprint-Bench 2 | 36.9% | — |
| Furniture Assembly | 58.3% | — |

## Multilingual

- GPT-6 Sol: 50.5 (#118)
- Qwen3.8 27B: 53.7 (#60)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Non-English | 1385 | 1430 |
| LMArena Chinese | 1405 | 1504 |
| LMArena French | 1410 | 1465 |
| LMArena German | 1390 | 1438 |
| LMArena Japanese | 1385 | 1384 |
| LMArena Korean | 1341 | 1393 |
| LMArena Russian | 1401 | 1415 |
| LMArena Spanish | 1384 | 1448 |

## Instruction Following

- GPT-6 Sol: 74.5 (#94)
- Qwen3.8 27B: 75.8 (#53)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Instruction Following | 1412 | 1439 |

## Long Context

- GPT-6 Sol: 43.1 (#108)
- Qwen3.8 27B: 44.3 (#70)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Longer Query | 1411 | 1450 |

## Writing & Preference

- GPT-6 Sol: 71.9 (#18)
- Qwen3.8 27B: 65.8 (#43)

| Benchmark | GPT-6 Sol | Qwen3.8 27B |
|---|---|---|
| LMArena Text | 1395 | 1441 |
| LMArena Creative Writing | 1378 | 1384 |
| EQ-Bench Creative Writing | 2125 | 1671 |
| LMArena Multi-Turn | 1412 | 1441 |

## FAQ

### Is GPT-6 Sol better than Qwen3.8 27B?

GPT-6 Sol is the stronger model overall, scoring 61.8 to 46.0 on the Noometry Index. Qwen3.8 27B costs 6.6× 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 Qwen3.8 27B?

Qwen3.8 27B is cheaper. It lists at $0.04 per million input tokens and $2.30 per million output tokens; GPT-6 Sol lists at $2 and $10.

### Is GPT-6 Sol or Qwen3.8 27B better for coding?

GPT-6 Sol scores higher on coding benchmarks: 60.1 versus 50.5 in the Noometry coding category.

### Which has the bigger context window?

GPT-6 Sol does, with 1.05M tokens against 262K.

### How many benchmarks do GPT-6 Sol and Qwen3.8 27B share?

30 benchmarks have published results for both models. GPT-6 Sol has 45 scored results on Noometry and Qwen3.8 27B has 31.