GPT-5 vs Qwen3 8
GPT-5 
Qwen3 8 Verdict
- Qwen3 8 (max) leads on quality: 63.3 vs 60.8.
- GPT-5 is stronger in instruction following, maths.
- Qwen3 8 (max) is stronger in agents & tools, coding, composite, human preference, knowledge, long context, multimodal, reasoning.
- Qwen3 8 (max) is 1.3× cheaper ($2.67 vs $3.44 per 1M blended).
- GPT-5 streams 2.2× faster (85 vs 38 tokens per second).
| Metric | GPT-5 | Qwen3 8 (max) |
|---|---|---|
| BenchLeader Index | 60.8 | 63.3 |
| Agents & tools score | 51.0 | 56.0 |
| Coding score | 60.9 | 66.5 |
| Composite score | 58.1 | 71.0 |
| Human preference score | 61.7 | 68.4 |
| Instruction following score | 65.0 | – |
| Knowledge score | 62.2 | 62.4 |
| Long context score | 65.6 | 65.7 |
| Maths score | 72.8 | 62.4 |
| Multimodal score | 59.6 | 67.3 |
| Reasoning score | 64.2 | 68.3 |
| Blended price $/M | $3.44 | $2.67 |
| Output speed | 85 tok/s | 38 tok/s |
| Time to first answer | 60.7 s | 55.5 s |
| Context window | 400k | 1M |
| Terminal-Bench | 49.6% | – |
| SciCode | 42.9% | 52.9% |
| Remote Labor Index | 1.7% | – |
| WeirdML | 39.8% | – |
| APEX-Agents | 18.3% | – |
| ProofBench | – | 58.0% |
| Epoch Capabilities Index | 150 | 156.6 |
| LMArena Text | 1427 | 1481 |
| LMArena Hard Prompts | 1449 | 1504 |
| LMArena Coding | 1462 | 1522 |
| LMArena WebDev | – | 1671 |
| LMArena Vision | 1232 | 1313 |
| LMArena Agent | – | 4 |
| LiveBench | – | 78.5% |
| LiveBench Reasoning | – | 88.2% |
| LiveBench Coding | – | 72.9% |
| LiveBench Agentic Coding | – | 64.7% |
| LiveBench Mathematics | – | 91.3% |
| LiveBench Data Analysis | – | 78.4% |
| LiveBench Language | – | 79.7% |
| AA Intelligence Index | 23.0 | 40.3 |
| IFBench | 73.1% | – |
| AA-LCR | 78.2% | 78.3% |
| MMMU-Pro | 74.2% | 82.3% |
| AA-Omniscience | -8.7 | 3.4 |
| Terminal-Bench Hard | 32.6% | – |
| GPQA Diamond (AA) | 85.3% | 92.7% |
| Humanity's Last Exam (AA) | 28.5% | 43.0% |
| SciCode (AA) | – | 53.2% |
| τ²-Bench Telecom (AA) | 84.8% | – |
| LiveCodeBench | – | 87.8% |
| MMLU-Pro | – | 88.6% |
| IOI | – | 68.9% |
| LegalBench | – | 83.6% |
| CorpFin | – | 65.8% |
| TaxEval | – | 75.5% |
| Terminal-Bench 2.1 (Vals) | – | 67.4% |
| SWE-bench (Vals) | – | 85.6% |
| GPQA Diamond (Vals) | – | 93.7% |
| Vals Index | – | 51.8 |
| PRBench Finance | 51.3% | – |
| PRBench Legal | 49.0% | – |
| VISTA | 49.7% | – |
| MultiNRC | 52.1% | – |
| TutorBench | 55.3% | – |
| Kagi LLM Benchmark | 72.7% | – |
| IFEval (HELM) | 87.5% | – |
| Omni-MATH (HELM) | 64.7% | – |
| WildBench (HELM) | 85.7% | – |
| MMLU-Pro (HELM) | 86.3% | – |
| GPQA Diamond (HELM) | 79.1% | – |
| HELM Capabilities mean | 80.7% | – |
| Aider Polyglot | 88.0% | – |
| SWE-bench Verified (any scaffold) | 75.6% | – |
Data as of 2026-09-12. Best configuration of each model; every score links to its source on the model pages.
GPT-5 vs Qwen3 8: questions
- Is GPT-5 better than Qwen3 8?
- Qwen3 8 (max) leads on quality: 63.3 vs 60.8. The BenchLeader Index combines every independent quality benchmark; Qwen3 8 (max) is ahead overall as of 2026-09-12, but check the category scores for your use.
- Is GPT-5 better than Qwen3 8 for coding?
- Qwen3 8 scores higher in coding (67 vs 61 on the category index, where 50 is average).
- Is GPT-5 better than Qwen3 8 for agentic tasks?
- Qwen3 8 scores higher in agentic tasks (56 vs 51 on the category index, where 50 is average).
- Which is cheaper, GPT-5 or Qwen3 8?
- Qwen3 8 is cheaper: $2.67 against $3.44 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5 or Qwen3 8?
- GPT-5 streams faster: 85 against 38 output tokens per second.
- Which has the larger context window?
- Qwen3 8 accepts more context: 1M against 400k tokens.