Model comparison
GPT-4 vs Grok-3 mini
Grok-3 mini is the stronger model overall, scoring 41.2 to 29.1 on the Noometry Index.
Last verified . 22 shared benchmarks.
Summary
- They share 22 benchmarks with published results for both. GPT-4 scores higher in 1 category and Grok-3 mini in 7 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in math, where Grok-3 mini leads 42.1 to 10.8.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 1.1% for GPT-4 and 77.8% for Grok-3 mini.
Side by side
| GPT-4 | Grok-3 mini | |
|---|---|---|
| Provider | OpenAI | xAI |
| Noometry Index | 29.1 | 41.2 |
| Released | 2023-03-14 | 2025-04-09 |
| Weights | Proprietary | Proprietary |
| Context window | 8K | — |
| Max output | 8K | — |
| Input $ / M tokens | $30 | — |
| Output $ / M tokens | $60 | — |
| Results tracked | 38 | 35 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Grok-3 mini leads
GPT-4: 31.6 (#283), Grok-3 mini: 40.8 (#131)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| WeirdML | 12.4% | 42.6% |
| LMArena Coding | 1254 | 1379 |
| Aider Polyglot | — | 49.3% |
| BigCodeBench Instruct | 46% | — |
| BigCodeBench Complete | 57.2% | — |
| HumanEval+ | 79.3% | — |
Agentic & Tool Use Not comparable
GPT-4: —, Grok-3 mini: —
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| METR Time Horizons | 36.1% | — |
Reasoning GPT-4 leads
GPT-4: 17.8 (#289), Grok-3 mini: 13.6 (#334)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| LMArena Hard Prompts | 1241 | 1375 |
| Epoch Capabilities Index | 125.89 | 140.35 |
| ARC-AGI-2 | — | 0.4% |
| Kagi LLM Benchmark | — | 61.3% |
| ARC-AGI-1 | — | 16.5% |
| Chess Puzzles | 4% | — |
| Mystery Game Puzzles | 12% | — |
| DTBench | 62.7% | — |
| LMCA | 17.1% | — |
| BIG-Bench Hard | 75.1% | — |
| ForecastBench | 57.8 | — |
| HellaSwag | 95.3% | — |
| WinoGrande | 87.5% | — |
Math Grok-3 mini leads
GPT-4: 10.8 (#309), Grok-3 mini: 42.1 (#85)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 1.1% | 77.8% |
| LMArena Math | 1269 | 1386 |
| MATH Level 5 | 23% | 90.9% |
| Omni-MATH | — | 31.8% |
| FrontierMath (Feb 2025 set) | — | 5.9% |
| GSM8K | 92% | — |
Knowledge Grok-3 mini leads
GPT-4: 18.4 (#282), Grok-3 mini: 46.4 (#81)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| GPQA Diamond | 35.7% | 76.3% |
| LMArena Expert | 1211 | 1395 |
| MMLU-Pro | — | 79.9% |
| Confabulations | — | 10.8% |
| GPQA (HELM) | — | 67.5% |
| MMLU | 86.4% | — |
| TriviaQA | 84.8% | — |
Multilingual Grok-3 mini leads
GPT-4: 40.6 (#215), Grok-3 mini: 48.1 (#145)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| LMArena Non-English | 1246 | 1352 |
| LMArena Chinese | 1242 | 1387 |
| LMArena French | 1283 | 1357 |
| LMArena German | 1251 | 1349 |
| LMArena Japanese | 1209 | 1342 |
| LMArena Korean | 1184 | 1335 |
| LMArena Russian | 1251 | 1353 |
| LMArena Spanish | 1261 | 1381 |
Instruction Following Grok-3 mini leads
GPT-4: 65.3 (#222), Grok-3 mini: 78.5 (#9)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| LMArena Instruction Following | 1241 | 1357 |
| IFEval | — | 95.1% |
Long Context Grok-3 mini leads
GPT-4: 37.7 (#212), Grok-3 mini: 41.0 (#147)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| LMArena Longer Query | 1244 | 1372 |
| Fiction.LiveBench | — | 66.7% |
Writing & Preference Grok-3 mini leads
GPT-4: 34.9 (#268), Grok-3 mini: 52.5 (#169)
| Benchmark | GPT-4 | Grok-3 mini |
|---|---|---|
| LMArena Text | 1263 | 1370 |
| LMArena Creative Writing | 1244 | 1342 |
| LMArena Multi-Turn | 1257 | 1355 |
| Short-Story Creative Writing | — | 73.5% |
| EQ-Bench Creative Writing | 752 | — |
| WildBench | — | 65.1% |
Frequently asked questions
Is GPT-4 better than Grok-3 mini?
Grok-3 mini is the stronger model overall, scoring 41.2 to 29.1 on the Noometry Index.
Is GPT-4 or Grok-3 mini better for coding?
Grok-3 mini scores higher on coding benchmarks: 40.8 versus 31.6 in the Noometry coding category.
How many benchmarks do GPT-4 and Grok-3 mini share?
22 benchmarks have published results for both models. GPT-4 has 38 scored results on Noometry and Grok-3 mini has 35.