# o3 vs Qwen3.8 Max

> Qwen3.8 Max is the stronger model overall, scoring 56.8 to 47.5 on the Noometry Index.

- Canonical page: https://noometry.com/compare/o3-vs-qwen3-8-max
- Last updated: 2026-10-10
- Shared benchmarks: 28

## Summary

- They share 28 benchmarks with published results for both. o3 scores higher in 2 categories and Qwen3.8 Max in 8 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Qwen3.8 Max leads 73.2 to 50.2.
- The biggest single-benchmark swing is FrontierMath (Tiers 1-3): 33.3% for o3 and 74.7% for Qwen3.8 Max.
- Qwen3.8 Max is cheaper at $2 / $6 per million input/output tokens, against $2 / $8 for o3.
- Qwen3.8 Max accepts more context: 1M tokens versus 200K.

## Snapshot

| | o3 | Qwen3.8 Max |
|---|---|---|
| Provider | OpenAI | Alibaba (Qwen) |
| Noometry Index | 47.5 | 56.8 |
| Rank | 61 | 22 |
| Context | 200K | 1M |
| Input $/M | $2 | $2 |
| Output $/M | $8 | $6 |
| Weights | Proprietary | Proprietary |

## Coding

- o3: 46.8 (#64)
- Qwen3.8 Max: 53.5 (#29)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Coding | 1408 | 1502 |
| SWE-bench Verified | 62.3% | — |
| DeepSWE | — | 57.5% |
| SWE-bench Verified (bash only) | 58.4% | — |
| Aider Polyglot | 81.3% | — |
| LMArena WebDev | — | 1674 |
| FrontierSWE | — | 17.8% |
| SciCode | — | 53.2% |
| GSO | 8.8% | — |
| WeirdML | 52.4% | — |
| CadEval | 74% | — |
| ALE-Bench | 933.55 | — |

## Agentic & Tool Use

- o3: 34.5 (#44)
- Qwen3.8 Max: 45.4 (#14)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| APEX-Agents | — | 63.3% |
| Berkeley Function Calling Leaderboard | 63% | — |
| GDPval | 30.8% | — |
| τ²-bench Banking | — | 55.1% |
| DeepResearch Bench | 45.2% | — |
| OSWorld | 23% | — |
| GDP.pdf | — | 23.2% |
| LMArena Search | 1144 | — |
| METR Time Horizons | 65.4% | — |

## Reasoning

- o3: 32.0 (#78)
- Qwen3.8 Max: 54.4 (#26)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| CritPt | 1.4% | 20% |
| Chess Puzzles | 38% | 40% |
| LMArena Hard Prompts | 1402 | 1496 |
| Mystery Game Puzzles | 29% | 38% |
| DTBench | 84.8% | 92% |
| LMCA | 39.7% | 46.2% |
| Epoch Capabilities Index | 146.86 | 156.41 |
| ARC-AGI-2 | 6.5% | — |
| SimpleBench | 53.1% | — |
| Kagi LLM Benchmark | 67.6% | — |
| NYT Connections (extended) | — | 88.3% |
| ARC-AGI-1 | 60.8% | — |
| EnigmaEval | 13.1% | — |
| ForecastBench | 62.5 | — |

## Math

- o3: 50.2 (#58)
- Qwen3.8 Max: 73.2 (#20)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| FrontierMath (Tiers 1-3) | 33.3% | 74.7% |
| OTIS Mock AIME 2024-2025 | 84.4% | 100% |
| LMArena Math | 1426 | 1499 |
| FrontierMath Tier 4 | — | 46.3% |
| ProofBench | — | 58% |
| Omni-MATH | 71.4% | — |
| MATH Level 5 | 97.8% | — |
| FrontierMath (Feb 2025 set) | 18.7% | — |
| FrontierMath Tier 4 (v1) | 2.1% | — |

## Knowledge

- o3: 54.6 (#52)
- Qwen3.8 Max: 61.7 (#27)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| GPQA Diamond | 81.8% | 92.7% |
| SimpleQA Verified | 49.4% | 47.3% |
| LMArena Expert | 1402 | 1507 |
| Humanity's Last Exam | 20.3% | — |
| MMLU-Pro | 85.9% | — |
| Confabulations | 14.4% | — |
| GPQA (HELM) | 75.3% | — |

## Multimodal

- o3: 41.4 (#36)
- Qwen3.8 Max: 37.2 (#75)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Vision | 1214 | 1314 |
| GeoBench | 74% | — |
| VPCT | 52% | — |
| Furniture Assembly | — | 20% |

## Multilingual

- o3: 51.7 (#105)
- Qwen3.8 Max: 56.7 (#18)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Non-English | 1401 | 1472 |
| LMArena Chinese | 1437 | 1538 |
| LMArena French | 1430 | 1503 |
| LMArena German | 1420 | 1483 |
| LMArena Japanese | 1403 | 1467 |
| LMArena Korean | 1370 | 1461 |
| LMArena Russian | 1406 | 1481 |
| LMArena Spanish | 1395 | 1492 |

## Instruction Following

- o3: 72.8 (#127)
- Qwen3.8 Max: 77.6 (#17)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Instruction Following | 1368 | 1479 |
| IFEval | 86.9% | — |

## Long Context

- o3: 53.3 (#6)
- Qwen3.8 Max: 45.6 (#31)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Longer Query | 1372 | 1489 |
| Fiction.LiveBench | 88.9% | — |
| CL-bench | 17.8% | — |

## Writing & Preference

- o3: 63.5 (#64)
- Qwen3.8 Max: 67.1 (#30)

| Benchmark | o3 | Qwen3.8 Max |
|---|---|---|
| LMArena Text | 1410 | 1483 |
| LMArena Creative Writing | 1359 | 1479 |
| LMArena Multi-Turn | 1405 | 1489 |
| Short-Story Creative Writing | 83.9% | — |
| EQ-Bench Creative Writing | 1676 | — |
| WildBench | 86.1% | — |

## FAQ

### Is o3 better than Qwen3.8 Max?

Qwen3.8 Max is the stronger model overall, scoring 56.8 to 47.5 on the Noometry Index.

### Which is cheaper, o3 or Qwen3.8 Max?

Qwen3.8 Max is cheaper. It lists at $2 per million input tokens and $6 per million output tokens; o3 lists at $2 and $8.

### Is o3 or Qwen3.8 Max better for coding?

Qwen3.8 Max scores higher on coding benchmarks: 53.5 versus 46.8 in the Noometry coding category.

### Which has the bigger context window?

Qwen3.8 Max does, with 1M tokens against 200K.

### How many benchmarks do o3 and Qwen3.8 Max share?

28 benchmarks have published results for both models. o3 has 63 scored results on Noometry and Qwen3.8 Max has 39.