Compare / Qwen3.6 Plus vs Qwen3 Coder 480B A35B
Qwen3.6 Plus vs Qwen3 Coder 480B A35B
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 | Qwen3.6 Plus | Qwen3 Coder 480B A35B |
|---|---|---|---|
| Arena (formerly LMArena)reported | OverallArena rating | 1,438 | 1,401 |
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,446 | 1,391 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,407 | 1,362 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,476 | 1,379 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,436 | 1,384 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,443 | 1,387 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,423 | 1,366 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 27.0 | 11.9 |
| Artificial Analysisreported | GPQA Diamond% of questions | 88.2% | 61.8% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 27.8% | 4.5% |
| Artificial Analysisreported | IFBench% of instructions | 75.2% | 40.5% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 78.3% | 45.7% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 43.9% | 18.9% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 97.7% | 43.6% |
| OpenHands Indexreported | Average% resolved | 52.9% | 30.9% |
| OpenHands Indexreported | Front end (SWE-Bench Multimodal)% resolved | 30.9% | 23.5% |
| OpenHands Indexreported | Greenfield (Commit0)% resolved | 25.0% | 0.0% |
| OpenHands Indexreported | Information gathering (GAIA)% resolved | 72.1% | 33.9% |
| OpenHands Indexreported | Issue resolution (SWE-Bench)% resolved | 74.2% | 62.4% |
| OpenHands Indexreported | Testing (SWT-Bench)% resolved | 62.1% | 34.9% |
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.