Claude Sonnet 5 vs GPT-5.2-Codex
Claude Sonnet 5 
GPT-5.2-Codex Verdict
- Claude Sonnet 5 (high) and GPT-5.2-Codex are level on quality (61.4 vs 60.8).
- Claude Sonnet 5 (high) is stronger in coding, composite, human preference, multimodal, reasoning.
- GPT-5.2-Codex is stronger in agents & tools, knowledge, long context, instruction following.
- Claude Sonnet 5 (high) is 1.2× cheaper ($4.00 vs $4.81 per 1M blended).
- Claude Sonnet 5 (high) streams 2.4× faster (60 vs 25 tokens per second).
| Metric | Claude Sonnet 5 (high) | GPT-5.2-Codex |
|---|---|---|
| BenchLeader Index | 61.4 | 60.8 |
| Agents & tools score | 61.8 | 63.5 |
| Coding score | 61.6 | 55.7 |
| Composite score | 69.5 | 56.8 |
| Human preference score | 66.2 | – |
| Knowledge score | 62.4 | 63.1 |
| Long context score | 64.8 | 67.8 |
| Multimodal score | 63.1 | 60.7 |
| Reasoning score | 66.8 | – |
| Instruction following score | – | 75.1 |
| Blended price $/M | $4.00 | $4.81 |
| Output speed | 60 tok/s | 25 tok/s |
| Time to first answer | 10.0 s | 6.7 s |
| Context window | 1M | 400k |
| Terminal-Bench | – | 66.5% |
| SciCode | 48.6% | – |
| WeirdML | 68.8% | – |
| APEX-Agents | – | 27.6% |
| LMArena Text | 1462 | – |
| LMArena Hard Prompts | 1490 | – |
| LMArena Coding | 1521 | – |
| LMArena WebDev | 1538 | 1339 |
| LMArena Vision | 1278 | – |
| LMArena Agent | 6.3 | – |
| LiveBench | – | 74.0% |
| LiveBench Reasoning | – | 77.7% |
| LiveBench Coding | – | 83.6% |
| LiveBench Agentic Coding | – | 49.4% |
| LiveBench Mathematics | – | 88.8% |
| LiveBench Data Analysis | – | 78.2% |
| LiveBench Language | – | 73.7% |
| AA Intelligence Index | 32.0 | 28.5 |
| IFBench | – | 77.6% |
| AA-LCR | 76.7% | 82.3% |
| MMMU-Pro | – | 76.3% |
| AA-Omniscience | -3.7 | -2.2 |
| Terminal-Bench Hard | – | 37.1% |
| GPQA Diamond (AA) | – | 89.9% |
| Humanity's Last Exam (AA) | 35.7% | 35.7% |
| SciCode (AA) | 54.3% | – |
| τ²-Bench Telecom (AA) | – | 92.1% |
| LiveCodeBench | – | 88.0% |
| SWE-Bench Pro | – | 41.0% |
| SWE-bench Verified (bash only) | – | 72.8% |
| SWE-bench Verified (any scaffold) | – | 72.8% |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
Claude Sonnet 5 vs GPT-5.2-Codex: questions
- Is Claude Sonnet 5 better than GPT-5.2-Codex?
- Claude Sonnet 5 (high) and GPT-5.2-Codex are level on quality (61.4 vs 60.8). The BenchLeader Index combines every independent quality benchmark; Claude Sonnet 5 (high) is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is Claude Sonnet 5 better than GPT-5.2-Codex for coding?
- Claude Sonnet 5 scores higher in coding (62 vs 56 on the category index, where 50 is average).
- Is Claude Sonnet 5 better than GPT-5.2-Codex for agentic tasks?
- GPT-5.2-Codex scores higher in agentic tasks (64 vs 62 on the category index, where 50 is average).
- Which is cheaper, Claude Sonnet 5 or GPT-5.2-Codex?
- Claude Sonnet 5 is cheaper: $4.00 against $4.81 per million tokens, blended at three input tokens per output token.
- Which is faster, Claude Sonnet 5 or GPT-5.2-Codex?
- Claude Sonnet 5 streams faster: 60 against 25 output tokens per second.
- Which has the larger context window?
- Claude Sonnet 5 accepts more context: 1M against 400k tokens.