Model comparison
GPT-5-Codex vs Grok 4
Grok 4 is the stronger model overall, scoring 48.1 to 37.9 on the Noometry Index.
Last verified . 3 shared benchmarks.
Summary
- They share 3 benchmarks with published results for both. GPT-5-Codex scores higher in 0 categories and Grok 4 in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in coding, where Grok 4 leads 50.3 to 42.4.
- The biggest single-benchmark swing is Terminal-Bench: 44.3% for GPT-5-Codex and 27.2% for Grok 4.
Side by side
| GPT-5-Codex | Grok 4 | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 37.9 | 48.1 |
| Released | 2025-09-15 | 2025-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | — |
| Max output | 128K | — |
| Input $ / M tokens | $1.25 | — |
| Output $ / M tokens | $10 | — |
| Results tracked | 3 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok 4 leads
GPT-5-Codex: 42.4 (#103), Grok 4: 50.3 (#46)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| WeirdML | 54.5% | 45.7% |
| Aider Polyglot | — | 79.6% |
| LMArena Coding | — | 1408 |
Agentic & Tool Use Grok 4 leads
GPT-5-Codex: 31.0 (#72), Grok 4: 32.3 (#68)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| Terminal-Bench | 44.3% | 27.2% |
| Berkeley Function Calling Leaderboard | — | 63% |
| GDPval | — | 21.1% |
| Cybench | — | 43% |
| DeepResearch Bench | — | 47.3% |
| BALROG | — | 43.6% |
| LMArena Search | — | 1142 |
| METR Time Horizons | — | 66.6% |
Reasoning Grok 4 leads
GPT-5-Codex: 30.9 (#83), Grok 4: 36.7 (#65)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| Kagi LLM Benchmark | 70.3% | 73.6% |
| ARC-AGI-2 | — | 16% |
| SimpleBench | — | 60.5% |
| ARC-AGI-1 | — | 66.7% |
| Chess Puzzles | — | 28% |
| LMArena Hard Prompts | — | 1409 |
| Epoch Capabilities Index | — | 146.44 |
| ForecastBench | — | 60.9 |
Math Not comparable
GPT-5-Codex: —, Grok 4: 48.4 (#64)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 84% |
| Omni-MATH | — | 60.3% |
| LMArena Math | — | 1422 |
| FrontierMath (Feb 2025 set) | — | 19.7% |
| FrontierMath Tier 4 (v1) | — | 2.1% |
Knowledge Not comparable
GPT-5-Codex: —, Grok 4: 53.8 (#55)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| GPQA Diamond | — | 87% |
| MMLU-Pro | — | 85.1% |
| Confabulations | — | 12.4% |
| GPQA (HELM) | — | 72.7% |
| LMArena Expert | — | 1415 |
Multimodal Not comparable
GPT-5-Codex: —, Grok 4: 33.7 (#94)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| LMArena Vision | — | 1210 |
| GeoBench | — | 45% |
Multilingual Not comparable
GPT-5-Codex: —, Grok 4: 51.8 (#103)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| LMArena Non-English | — | 1403 |
| LMArena Chinese | — | 1427 |
| LMArena French | — | 1418 |
| LMArena German | — | 1429 |
| LMArena Japanese | — | 1394 |
| LMArena Korean | — | 1377 |
| LMArena Russian | — | 1410 |
| LMArena Spanish | — | 1420 |
Instruction Following Not comparable
GPT-5-Codex: —, Grok 4: 79.2 (#5)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| IFEval | — | 94.9% |
| LMArena Instruction Following | — | 1387 |
Long Context Not comparable
GPT-5-Codex: —, Grok 4: 63.1 (#4)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| Fiction.LiveBench | — | 94.4% |
| LMArena Longer Query | — | 1409 |
Writing & Preference Not comparable
GPT-5-Codex: —, Grok 4: 58.5 (#116)
| Benchmark | GPT-5-Codex | Grok 4 |
|---|---|---|
| LMArena Text | — | 1411 |
| LMArena Creative Writing | — | 1397 |
| Short-Story Creative Writing | — | 76.9% |
| WildBench | — | 79.7% |
| LMArena Multi-Turn | — | 1416 |
Frequently asked questions
Is GPT-5-Codex better than Grok 4?
Grok 4 is the stronger model overall, scoring 48.1 to 37.9 on the Noometry Index.
Is GPT-5-Codex or Grok 4 better for coding?
Grok 4 scores higher on coding benchmarks: 50.3 versus 42.4 in the Noometry coding category.
How many benchmarks do GPT-5-Codex and Grok 4 share?
3 benchmarks have published results for both models. GPT-5-Codex has 3 scored results on Noometry and Grok 4 has 48.