GPT-5 vs Qwen3 7
GPT-5 
Qwen3 7 Verdict
- GPT-5 and Qwen3 7 (max) are level on quality (60.8 vs 61.2).
- GPT-5 is stronger in composite, human preference, maths, multimodal.
- Qwen3 7 (max) is stronger in agents & tools, instruction following, knowledge, long context, reasoning.
- They cost about the same ($3.44 per 1M blended).
- Qwen3 7 (max) streams 1.7× faster (142 vs 85 tokens per second).
| Metric | GPT-5 | Qwen3 7 (max) |
|---|---|---|
| BenchLeader Index | 60.8 | 61.2 |
| Agents & tools score | 51.0 | 55.8 |
| Coding score | 60.9 | 60.9 |
| Composite score | 58.1 | 56.4 |
| Human preference score | 61.7 | 57.1 |
| Instruction following score | 65.0 | 77.6 |
| Knowledge score | 62.2 | 63.6 |
| Long context score | 65.6 | 66.0 |
| Maths score | 72.8 | 57.4 |
| Multimodal score | 59.6 | – |
| Reasoning score | 64.2 | 66.8 |
| Blended price $/M | $3.44 | $3.75 |
| Output speed | 85 tok/s | 142 tok/s |
| Time to first answer | 60.7 s | 19.6 s |
| Context window | 400k | 1M |
| GPQA Diamond | – | 90.9% |
| FrontierMath Tiers 1–3 | – | 64.6% |
| FrontierMath Tier 4 | – | 34.1% |
| OTIS Mock AIME | – | 95.6% |
| SWE-bench Verified (Epoch) | – | 77.3% |
| SimpleQA Verified | – | 55.8% |
| Terminal-Bench | 49.6% | – |
| SimpleBench | – | 70.4% |
| SciCode | 42.9% | 48.8% |
| Remote Labor Index | 1.7% | – |
| WeirdML | 39.8% | – |
| APEX-Agents | 18.3% | – |
| ProofBench | – | 26.0% |
| Epoch Capabilities Index | 150 | 153.7 |
| LMArena Text | 1427 | 1474 |
| LMArena Hard Prompts | 1449 | 1495 |
| LMArena Coding | 1462 | 1525 |
| LMArena WebDev | – | 1517 |
| LMArena Vision | 1232 | – |
| LMArena Agent | – | -3.1 |
| LiveBench | – | 73.1% |
| LiveBench Reasoning | – | 83.3% |
| LiveBench Coding | – | 74.2% |
| LiveBench Agentic Coding | – | 43.6% |
| LiveBench Mathematics | – | 85.3% |
| LiveBench Data Analysis | – | 71.8% |
| LiveBench Language | – | 79.7% |
| AA Intelligence Index | 23.0 | 29.9 |
| IFBench | 73.1% | 80.5% |
| AA-LCR | 78.2% | 79.0% |
| MMMU-Pro | 74.2% | – |
| AA-Omniscience | -8.7 | 13.5 |
| Terminal-Bench Hard | 32.6% | 50.8% |
| GPQA Diamond (AA) | 85.3% | 92.3% |
| Humanity's Last Exam (AA) | 28.5% | 40.5% |
| SciCode (AA) | – | 49.5% |
| τ²-Bench Telecom (AA) | 84.8% | 94.7% |
| LiveCodeBench | – | 87.1% |
| MMLU-Pro | – | 89.3% |
| LegalBench | – | 84.9% |
| CorpFin | – | 63.7% |
| TaxEval | – | 75.3% |
| Terminal-Bench 2.1 (Vals) | – | 61.0% |
| SWE-bench (Vals) | – | 68.8% |
| GPQA Diamond (Vals) | – | 90.2% |
| Vals Index | – | 44.8 |
| PRBench Finance | 51.3% | – |
| PRBench Legal | 49.0% | – |
| VISTA | 49.7% | – |
| MultiNRC | 52.1% | – |
| TutorBench | 55.3% | – |
| EQ-Bench 4 | – | 1110 |
| 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 7: questions
- Is GPT-5 better than Qwen3 7?
- GPT-5 and Qwen3 7 (max) are level on quality (60.8 vs 61.2). The BenchLeader Index combines every independent quality benchmark; Qwen3 7 (max) is ahead overall as of 2026-09-12, but check the category scores for your use.
- Is GPT-5 better than Qwen3 7 for coding?
- They are level in coding (61 each).
- Is GPT-5 better than Qwen3 7 for agentic tasks?
- Qwen3 7 scores higher in agentic tasks (56 vs 51 on the category index, where 50 is average).
- Which is cheaper, GPT-5 or Qwen3 7?
- GPT-5 is cheaper: $3.44 against $3.75 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5 or Qwen3 7?
- Qwen3 7 streams faster: 142 against 85 output tokens per second.
- Which has the larger context window?
- Qwen3 7 accepts more context: 1M against 400k tokens.