Gemini 3 Pro vs GPT-5.2-Codex
Gemini 3 Pro 
GPT-5.2-Codex Verdict
- Gemini 3 Pro and GPT-5.2-Codex are level on quality (61.7 vs 60.8).
- Gemini 3 Pro is stronger in coding, composite, human preference, maths, multimodal, reasoning.
- GPT-5.2-Codex is stronger in agents & tools, instruction following, knowledge, long context.
- They cost about the same ($4.50 per 1M blended).
| Metric | Gemini 3 Pro | GPT-5.2-Codex |
|---|---|---|
| BenchLeader Index | 61.7 | 60.8 |
| Agents & tools score | 62.3 | 63.5 |
| Coding score | 58.9 | 55.7 |
| Composite score | 64.3 | 56.8 |
| Human preference score | 69.0 | – |
| Instruction following score | 64.3 | 75.1 |
| Knowledge score | 60.9 | 63.1 |
| Long context score | 64.5 | 67.8 |
| Maths score | 48.4 | – |
| Multimodal score | 65.7 | 60.7 |
| Reasoning score | 67.1 | – |
| Blended price $/M | $4.50 | $4.81 |
| Output speed | – | 25 tok/s |
| Time to first answer | – | 6.7 s |
| Context window | 1M | 400k |
| GPQA Diamond | 92.6% | – |
| OTIS Mock AIME | 91.4% | – |
| SWE-bench Verified (Epoch) | 72.9% | – |
| Humanity's Last Exam | 37.5% | – |
| Terminal-Bench | 69.4% | 66.5% |
| SimpleBench | 76.4% | – |
| GDPval | 40.3% | – |
| Remote Labor Index | 1.3% | – |
| WeirdML | 69.9% | – |
| APEX-Agents | 31.5% | 27.6% |
| ProofBench | 20.0% | – |
| GSO-Bench | 18.6% | – |
| Epoch Capabilities Index | 153.0 | – |
| LMArena Text | 1486 | – |
| LMArena Hard Prompts | 1504 | – |
| LMArena Coding | 1518 | – |
| LMArena WebDev | 1439 | 1339 |
| LMArena Vision | 1305 | – |
| 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 | 28.0 | 28.5 |
| IFBench | 70.4% | 77.6% |
| AA-LCR | 76.0% | 82.3% |
| MMMU-Pro | 80.2% | 76.3% |
| AA-Omniscience | 15.3 | -2.2 |
| Terminal-Bench Hard | 41.7% | 37.1% |
| GPQA Diamond (AA) | 90.8% | 89.9% |
| Humanity's Last Exam (AA) | 39.7% | 35.7% |
| τ²-Bench Telecom (AA) | 87.1% | 92.1% |
| LiveCodeBench | 86.4% | 88.0% |
| MMLU-Pro | 90.1% | – |
| AIME 2026 | 91.7% | – |
| HMMT February 2026 | 86.4% | – |
| MathArena Apex | 23.4% | – |
| SWE-Bench Pro | 43.3% | 41.0% |
| MCP Atlas | 70.3% | – |
| MultiChallenge | 65.7% | – |
| PRBench Finance | 39.2% | – |
| PRBench Legal | 40.6% | – |
| VISTA | 51.5% | – |
| MultiNRC | 59.0% | – |
| TutorBench | 53.7% | – |
| MMMU-Pro (official) | 81.0% | – |
| Kagi LLM Benchmark | 80.1% | – |
| SWE-bench Verified (bash only) | 74.2% | 72.8% |
| SWE-bench Verified (any scaffold) | 77.4% | 72.8% |
| ARC-AGI-1 | 75.0% | – |
| ARC-AGI-2 | 54.0% | – |
| BFCL Overall | 72.5% | – |
Data as of 2026-09-10. Best configuration of each model; every score links to its source on the model pages.
Gemini 3 Pro vs GPT-5.2-Codex: questions
- Is Gemini 3 Pro better than GPT-5.2-Codex?
- Gemini 3 Pro and GPT-5.2-Codex are level on quality (61.7 vs 60.8). The BenchLeader Index combines every independent quality benchmark; Gemini 3 Pro is ahead overall as of 2026-09-10, but check the category scores for your use.
- Is Gemini 3 Pro better than GPT-5.2-Codex for coding?
- Gemini 3 Pro scores higher in coding (59 vs 56 on the category index, where 50 is average).
- Is Gemini 3 Pro 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, Gemini 3 Pro or GPT-5.2-Codex?
- Gemini 3 Pro is cheaper: $4.50 against $4.81 per million tokens, blended at three input tokens per output token.
- Which has the larger context window?
- Gemini 3 Pro accepts more context: 1M against 400k tokens.