GPT-5.2 vs GPT-5.2-Codex
GPT-5.2 Verdict
- GPT-5.2 and GPT-5.2-Codex are level on quality (61.6 vs 60.8).
- GPT-5.2 is stronger in composite, human preference, long context, reasoning.
- GPT-5.2-Codex is stronger in agents & tools, coding, instruction following, knowledge, multimodal.
- They cost about the same ($4.81 per 1M blended).
- GPT-5.2 streams 2.7× faster (68 vs 25 tokens per second).
| Metric | GPT-5.2 | GPT-5.2-Codex |
|---|---|---|
| BenchLeader Index | 61.6 | 60.8 |
| Agents & tools score | 61.2 | 63.5 |
| Coding score | 53.7 | 55.7 |
| Composite score | 67.5 | 56.8 |
| Human preference score | 67.9 | – |
| Instruction following score | 67.4 | 75.1 |
| Knowledge score | 61.1 | 63.1 |
| Long context score | 67.9 | 67.8 |
| Multimodal score | 59.9 | 60.7 |
| Reasoning score | 61.3 | – |
| Blended price $/M | $4.81 | $4.81 |
| Output speed | 68 tok/s | 25 tok/s |
| Time to first answer | 129.4 s | 6.7 s |
| Context window | 400k | 400k |
| Humanity's Last Exam | 27.8% | – |
| Terminal-Bench | 64.9% | 66.5% |
| SimpleBench | 45.8% | – |
| Remote Labor Index | 2.1% | – |
| APEX-Agents | 23.0% | 27.6% |
| Epoch Capabilities Index | 153.5 | – |
| LMArena Text | 1476 | – |
| LMArena Hard Prompts | 1497 | – |
| LMArena Coding | 1515 | – |
| LMArena WebDev | 1417 | 1339 |
| LMArena Vision | 1268 | – |
| 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 | 30.4 | 28.5 |
| IFBench | 75.4% | 77.6% |
| AA-LCR | 82.7% | 82.3% |
| MMMU-Pro | – | 76.3% |
| AA-Omniscience | -0.9 | -2.2 |
| Terminal-Bench Hard | 47.0% | 37.1% |
| GPQA Diamond (AA) | 90.3% | 89.9% |
| Humanity's Last Exam (AA) | 37.7% | 35.7% |
| τ²-Bench Telecom (AA) | 84.8% | 92.1% |
| LiveCodeBench | – | 88.0% |
| SWE-Bench Pro | 29.9% | 41.0% |
| VISTA | 46.6% | – |
| MultiNRC | 42.2% | – |
| TutorBench | 53.5% | – |
| Kagi LLM Benchmark | 73.3% | – |
| SWE-bench Verified (bash only) | 69.0% | 72.8% |
| SWE-bench Verified (any scaffold) | – | 72.8% |
| ARC-AGI-1 | 94.5% | – |
| ARC-AGI-2 | 72.9% | – |
| BFCL Overall | 55.9% | – |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
GPT-5.2 vs GPT-5.2-Codex: questions
- Is GPT-5.2 better than GPT-5.2-Codex?
- GPT-5.2 and GPT-5.2-Codex are level on quality (61.6 vs 60.8). The BenchLeader Index combines every independent quality benchmark; GPT-5.2 is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is GPT-5.2 better than GPT-5.2-Codex for coding?
- GPT-5.2-Codex scores higher in coding (56 vs 54 on the category index, where 50 is average).
- Is GPT-5.2 better than GPT-5.2-Codex for agentic tasks?
- GPT-5.2-Codex scores higher in agentic tasks (64 vs 61 on the category index, where 50 is average).
- Which is cheaper, GPT-5.2 or GPT-5.2-Codex?
- GPT-5.2 is cheaper: $4.81 against $4.81 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.2 or GPT-5.2-Codex?
- GPT-5.2 streams faster: 68 against 25 output tokens per second.
- Which has the larger context window?
- Both accept 400k tokens of context.