Compare / Qwen3 Coder 480B A35B vs Claude Opus 4.8
Qwen3 Coder 480B A35B vs Claude Opus 4.8
20 results from 3 sources that measured both models. Each row is in the source's own units; there is no overall winner.
| Benchmark | Metric | Qwen3 Coder 480B A35B | Claude Opus 4.8 |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,401 | 1,469 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,391 | 1,490 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,362 | 1,467 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,379 | 1,525 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,384 | 1,491 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,387 | 1,480 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,366 | 1,474 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 11.9 | 41.8 |
| Artificial Analysisreported | GPQA Diamond% of questions | 61.8% | 92.0% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 4.5% | 48.7% |
| Artificial Analysisreported | IFBench% of instructions | 40.5% | 62.2% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 45.7% | 77.7% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 18.9% | 58.3% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 43.6% | 94.4% |
| OpenHands Indexreported | Average% resolved | 30.9% | 71.9% |
| OpenHands Indexreported | Front end (SWE-Bench Multimodal)% resolved | 23.5% | 50.0% |
| OpenHands Indexreported | Greenfield (Commit0)% resolved | 0.0% | 62.5% |
| OpenHands Indexreported | Information gathering (GAIA)% resolved | 33.9% | 78.8% |
| OpenHands Indexreported | Issue resolution (SWE-Bench)% resolved | 62.4% | 83.8% |
| OpenHands Indexreported | Testing (SWT-Bench)% resolved | 34.9% | 84.3% |
Bold green marks the better value on that metric. Where a source reports ranges that overlap, the difference may not be meaningful; see the benchmark page for ranges.