GPT-5 vs o3
GPT-5 Verdict
- GPT-5 and o3 are level on quality (60.8 vs 61.6).
- GPT-5 is stronger in composite, knowledge, long context, multimodal, reasoning.
- o3 is stronger in agents & tools, coding, human preference, maths.
- They cost about the same ($3.44 per 1M blended).
| Metric | GPT-5 | o3 |
|---|---|---|
| BenchLeader Index | 60.8 | 61.6 |
| Agents & tools score | 51.0 | 70.3 |
| Coding score | 60.9 | 62.0 |
| Composite score | 58.1 | 54.5 |
| Human preference score | 61.7 | 62.3 |
| Instruction following score | 65.0 | 65.0 |
| Knowledge score | 62.2 | 58.0 |
| Long context score | 65.6 | 63.8 |
| Maths score | 72.8 | 78.6 |
| Multimodal score | 59.6 | 57.9 |
| Reasoning score | 64.2 | 61.5 |
| Blended price $/M | $3.44 | $3.50 |
| Output speed | 85 tok/s | 107 tok/s |
| Time to first answer | 60.7 s | 6.3 s |
| Context window | 400k | 200k |
| Terminal-Bench | 49.6% | – |
| SciCode | 42.9% | – |
| Remote Labor Index | 1.7% | – |
| WeirdML | 39.8% | – |
| APEX-Agents | 18.3% | – |
| Epoch Capabilities Index | 150 | 146.9 |
| LMArena Text | 1427 | 1432 |
| LMArena Hard Prompts | 1449 | 1441 |
| LMArena Coding | 1462 | 1460 |
| LMArena Vision | 1232 | 1214 |
| AA Intelligence Index | 23.0 | 20.2 |
| IFBench | 73.1% | 71.4% |
| AA-LCR | 78.2% | 74.7% |
| MMMU-Pro | 74.2% | 70.1% |
| AA-Omniscience | -8.7 | -15.6 |
| Terminal-Bench Hard | 32.6% | 37.1% |
| GPQA Diamond (AA) | 85.3% | 82.7% |
| Humanity's Last Exam (AA) | 28.5% | 20.1% |
| τ²-Bench Telecom (AA) | 84.8% | 80.7% |
| PRBench Finance | 51.3% | 47.7% |
| PRBench Legal | 49.0% | 48.6% |
| VISTA | 49.7% | – |
| MultiNRC | 52.1% | – |
| TutorBench | 55.3% | – |
| MMMU (validation) | – | 82.9% |
| MMMU-Pro (official) | – | 76.4% |
| Kagi LLM Benchmark | 72.7% | 67.6% |
| IFEval (HELM) | 87.5% | 86.9% |
| Omni-MATH (HELM) | 64.7% | 71.4% |
| WildBench (HELM) | 85.7% | 86.1% |
| MMLU-Pro (HELM) | 86.3% | 85.9% |
| GPQA Diamond (HELM) | 79.1% | 75.3% |
| HELM Capabilities mean | 80.7% | 81.1% |
| Aider Polyglot | 88.0% | 81.3% |
| SWE-bench Verified (bash only) | – | 58.4% |
| SWE-bench Verified (any scaffold) | 75.6% | 58.4% |
| BFCL Overall | – | 63.0% |
Data as of 2026-09-12. Best configuration of each model; every score links to its source on the model pages.
GPT-5 vs o3: questions
- Is GPT-5 better than o3?
- GPT-5 and o3 are level on quality (60.8 vs 61.6). The BenchLeader Index combines every independent quality benchmark; o3 is ahead overall as of 2026-09-12, but check the category scores for your use.
- Is GPT-5 better than o3 for coding?
- o3 scores higher in coding (62 vs 61 on the category index, where 50 is average).
- Is GPT-5 better than o3 for agentic tasks?
- o3 scores higher in agentic tasks (70 vs 51 on the category index, where 50 is average).
- Which is cheaper, GPT-5 or o3?
- GPT-5 is cheaper: $3.44 against $3.50 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5 or o3?
- o3 streams faster: 107 against 85 output tokens per second.
- Which has the larger context window?
- GPT-5 accepts more context: 400k against 200k tokens.