Model comparison
GPT-5.2 Codex vs o3
o3 is the stronger model overall, scoring 47.5 to 42.6 on the Noometry Index.
Last verified . 2 shared benchmarks.
Summary
- They share 2 benchmarks with published results for both. GPT-5.2 Codex scores higher in 1 category and o3 in 1 category; 2 gaps are clear of the uncertainty.
- The widest gap is in agentic & tool use, where GPT-5.2 Codex leads 41.0 to 34.5.
- The biggest single-benchmark swing is SWE-bench Verified (bash only): 72.8% for GPT-5.2 Codex and 58.4% for o3.
- o3 is cheaper at $2 / $8 per million input/output tokens, against $1.75 / $14 for GPT-5.2 Codex.
- GPT-5.2 Codex accepts more context: 400K tokens versus 200K.
Side by side
| GPT-5.2 Codex | o3 | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 42.6 | 47.5 |
| Released | 2025-12-18 | 2025-04-16 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 200K |
| Max output | 128K | 100K |
| Input $ / M tokens | $1.75 | $2 |
| Output $ / M tokens | $14 | $8 |
| Results tracked | 5 | 63 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
GPT-5.2 Codex: 45.5 (#71), o3: 46.8 (#64)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| SWE-bench Verified (bash only) | 72.8% | 58.4% |
| ALE-Bench | 1,300 | 933.55 |
| SWE-bench Verified | — | 62.3% |
| Aider Polyglot | — | 81.3% |
| LMArena WebDev | 1339 | — |
| SWE-bench Multilingual | 66.3% | — |
| GSO | — | 8.8% |
| WeirdML | — | 52.4% |
| LMArena Coding | — | 1408 |
| CadEval | — | 74% |
Agentic & Tool Use GPT-5.2 Codex leads
GPT-5.2 Codex: 41.0 (#22), o3: 34.5 (#44)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| Terminal-Bench | 66.5% | — |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 30.8% |
| DeepResearch Bench | — | 45.2% |
| OSWorld | — | 23% |
| LMArena Search | — | 1144 |
| METR Time Horizons | — | 65.4% |
Reasoning Not comparable
GPT-5.2 Codex: —, o3: 32.0 (#78)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| ARC-AGI-2 | — | 6.5% |
| SimpleBench | — | 53.1% |
| Kagi LLM Benchmark | — | 67.6% |
| ARC-AGI-1 | — | 60.8% |
| CritPt | — | 1.4% |
| Chess Puzzles | — | 38% |
| EnigmaEval | — | 13.1% |
| LMArena Hard Prompts | — | 1402 |
| Mystery Game Puzzles | — | 29% |
| DTBench | — | 84.8% |
| LMCA | — | 39.7% |
| Epoch Capabilities Index | — | 146.86 |
| ForecastBench | — | 62.5 |
Math Not comparable
GPT-5.2 Codex: —, o3: 50.2 (#58)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| FrontierMath (Tiers 1-3) | — | 33.3% |
| OTIS Mock AIME 2024-2025 | — | 84.4% |
| Omni-MATH | — | 71.4% |
| LMArena Math | — | 1426 |
| MATH Level 5 | — | 97.8% |
| FrontierMath (Feb 2025 set) | — | 18.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge Not comparable
GPT-5.2 Codex: —, o3: 54.6 (#52)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| GPQA Diamond | — | 81.8% |
| Humanity's Last Exam | — | 20.3% |
| SimpleQA Verified | — | 49.4% |
| MMLU-Pro | — | 85.9% |
| Confabulations | — | 14.4% |
| GPQA (HELM) | — | 75.3% |
| LMArena Expert | — | 1402 |
Multimodal Not comparable
GPT-5.2 Codex: —, o3: 41.4 (#36)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| LMArena Vision | — | 1214 |
| GeoBench | — | 74% |
| VPCT | — | 52% |
Multilingual Not comparable
GPT-5.2 Codex: —, o3: 51.7 (#105)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| LMArena Non-English | — | 1401 |
| LMArena Chinese | — | 1437 |
| LMArena French | — | 1430 |
| LMArena German | — | 1420 |
| LMArena Japanese | — | 1403 |
| LMArena Korean | — | 1370 |
| LMArena Russian | — | 1406 |
| LMArena Spanish | — | 1395 |
Instruction Following Not comparable
GPT-5.2 Codex: —, o3: 72.8 (#127)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| IFEval | — | 86.9% |
| LMArena Instruction Following | — | 1368 |
Long Context Not comparable
GPT-5.2 Codex: —, o3: 53.3 (#6)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| Fiction.LiveBench | — | 88.9% |
| CL-bench | — | 17.8% |
| LMArena Longer Query | — | 1372 |
Writing & Preference Not comparable
GPT-5.2 Codex: —, o3: 63.5 (#64)
| Benchmark | GPT-5.2 Codex | o3 |
|---|---|---|
| LMArena Text | — | 1410 |
| LMArena Creative Writing | — | 1359 |
| Short-Story Creative Writing | — | 83.9% |
| EQ-Bench Creative Writing | — | 1676 |
| WildBench | — | 86.1% |
| LMArena Multi-Turn | — | 1405 |
Frequently asked questions
Is GPT-5.2 Codex better than o3?
o3 is the stronger model overall, scoring 47.5 to 42.6 on the Noometry Index.
Which is cheaper, GPT-5.2 Codex or o3?
o3 is cheaper. It lists at $2 per million input tokens and $8 per million output tokens; GPT-5.2 Codex lists at $1.75 and $14.
Is GPT-5.2 Codex or o3 better for coding?
o3 scores higher on coding benchmarks: 46.8 versus 45.5 in the Noometry coding category.
Which has the bigger context window?
GPT-5.2 Codex does, with 400K tokens against 200K.
How many benchmarks do GPT-5.2 Codex and o3 share?
2 benchmarks have published results for both models. GPT-5.2 Codex has 5 scored results on Noometry and o3 has 63.