Compare / Qwen3 Next 80B A3B Thinking vs GPT-5.6 Terra
Qwen3 Next 80B A3B Thinking vs GPT-5.6 Terra
19 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 Next 80B A3B Thinking | GPT-5.6 Terra |
|---|---|---|---|
| Arena (formerly LMArena)reported | Business, management and financeArena rating | 1,367 | 1,466 |
| Arena (formerly LMArena)reported | Creative writingArena rating | 1,326 | 1,428 |
| Arena (formerly LMArena)reported | Expert promptsArena rating | 1,394 | 1,504 |
| Arena (formerly LMArena)reported | Instruction followingArena rating | 1,359 | 1,462 |
| Arena (formerly LMArena)reported | OverallArena rating | 1,369 | 1,465 |
| Arena (formerly LMArena)reported | Writing, literature and languageArena rating | 1,341 | 1,445 |
| Artificial Analysisreported | Artificial Analysis Coding Indexindex score | 17.4 | 76.7 |
| Artificial Analysisreported | Artificial Analysis Intelligence Indexindex score | 11.2 | 42.1 |
| Artificial Analysisreported | GPQA Diamond% of questions | 75.9% | 92.5% |
| Artificial Analysisreported | Humanity's Last Exam% of questions | 12.6% | 42.9% |
| Artificial Analysisreported | IFBench% of instructions | 60.7% | 71.2% |
| Artificial Analysisreported | Long-context reasoning (AA-LCR)% of questions | 63.7% | 83.0% |
| Artificial Analysisreported | Terminal-Bench 2.1% of tasks | 6.7% | 88.0% |
| Artificial Analysisreported | Terminal-Bench Hard% of tasks | 9.8% | 62.9% |
| Artificial Analysisreported | Τ-bench banking% of tasks | 6.2% | 40.2% |
| Artificial Analysisreported | Τ²-bench telecom% of tasks | 41.5% | 86.3% |
| UGI Leaderboardreported | Requested-length error% off the requested word count | 67.0% | 22.0% |
| UGI Leaderboardreported | Style adherencescore from 0 to 1 | 0.28 | 0.39 |
| UGI Leaderboardreported | Writing scorescore out of 100 | 27.2 | 60.4 |
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.