GPT-5.2-Codex vs GPT-5.4
GPT-5.2-Codex Verdict
- GPT-5.4 leads on quality: 63.5 vs 60.8.
- GPT-5.2-Codex is stronger in coding, instruction following, long context.
- GPT-5.4 is stronger in agents & tools, composite, knowledge, multimodal, human preference, reasoning.
- GPT-5.2-Codex is 1.2× cheaper ($4.81 vs $5.63 per 1M blended).
- GPT-5.4 streams 5.2× faster (129 vs 25 tokens per second).
| Metric | GPT-5.2-Codex | GPT-5.4 |
|---|---|---|
| BenchLeader Index | 60.8 | 63.5 |
| Agents & tools score | 63.5 | 66.2 |
| Coding score | 55.7 | 51.6 |
| Composite score | 56.8 | 78.3 |
| Instruction following score | 75.1 | 71.9 |
| Knowledge score | 63.1 | 66.9 |
| Long context score | 67.8 | 67.6 |
| Multimodal score | 60.7 | 64.0 |
| Human preference score | – | 64.7 |
| Reasoning score | – | 62.6 |
| Blended price $/M | $4.81 | $5.63 |
| Output speed | 25 tok/s | 129 tok/s |
| Time to first answer | 6.7 s | 124.6 s |
| Context window | 400k | 1.1M |
| Terminal-Bench | 66.5% | 81.8% |
| APEX-Agents | 27.6% | 34.9% |
| Epoch Capabilities Index | – | 156.9 |
| LMArena Text | – | 1466 |
| LMArena Hard Prompts | – | 1488 |
| LMArena Coding | – | 1514 |
| LMArena WebDev | 1339 | 1386 |
| LMArena Vision | – | 1293 |
| 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 | 28.5 | 39.0 |
| IFBench | 77.6% | 74.0% |
| AA-LCR | 82.3% | 82.0% |
| MMMU-Pro | 76.3% | 78.4% |
| AA-Omniscience | -2.2 | 5.8 |
| Terminal-Bench Hard | 37.1% | 57.6% |
| GPQA Diamond (AA) | 89.9% | 92.0% |
| Humanity's Last Exam (AA) | 35.7% | 43.7% |
| τ²-Bench Telecom (AA) | 92.1% | 87.1% |
| LiveCodeBench | 88.0% | – |
| SWE-Bench Pro | 41.0% | – |
| HiL-Bench | – | 9.7% |
| EQ-Bench 4 | – | 1272 |
| Kagi LLM Benchmark | – | 63.8% |
| 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.
GPT-5.2-Codex vs GPT-5.4: questions
- Is GPT-5.2-Codex better than GPT-5.4?
- GPT-5.4 leads on quality: 63.5 vs 60.8. The BenchLeader Index combines every independent quality benchmark; GPT-5.4 is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is GPT-5.2-Codex better than GPT-5.4 for coding?
- GPT-5.2-Codex scores higher in coding (56 vs 52 on the category index, where 50 is average).
- Is GPT-5.2-Codex better than GPT-5.4 for agentic tasks?
- GPT-5.4 scores higher in agentic tasks (66 vs 64 on the category index, where 50 is average).
- Which is cheaper, GPT-5.2-Codex or GPT-5.4?
- GPT-5.2-Codex is cheaper: $4.81 against $5.63 per million tokens, blended at three input tokens per output token.
- Which is faster, GPT-5.2-Codex or GPT-5.4?
- GPT-5.4 streams faster: 129 against 25 output tokens per second.
- Which has the larger context window?
- GPT-5.4 accepts more context: 1.1M against 400k tokens.