Model comparison
Grok-2 (Dec 2024) vs Grok 4 Fast
Grok 4 Fast is the stronger model overall, scoring 39.4 to 33.7 on the Noometry Index.
Last verified . 20 shared benchmarks.
Summary
- They share 20 benchmarks with published results for both. Grok-2 (Dec 2024) scores higher in 1 category and Grok 4 Fast in 7 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in long context, where Grok 4 Fast leads 63.2 to 38.8.
- The biggest single-benchmark swing is WeirdML: 22.2% for Grok-2 (Dec 2024) and 42.9% for Grok 4 Fast.
Side by side
| Grok-2 (Dec 2024) | Grok 4 Fast | |
|---|---|---|
| Provider | xAI | xAI |
| Noometry Index | 33.7 | 39.4 |
| Released | 2024-08-13 | 2025-09-19 |
| Weights | Proprietary | Proprietary |
| Context window | — | — |
| Max output | — | — |
| Input $ / M tokens | — | — |
| Output $ / M tokens | — | — |
| Results tracked | 34 | 30 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
Grok-2 (Dec 2024): 33.3 (#258), Grok 4 Fast: 32.5 (#271)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| WeirdML | 22.2% | 42.9% |
| LMArena Coding | 1287 | 1429 |
| LMArena WebDev | — | 1159 |
| LiveBench Coding | 46.4% | — |
Agentic & Tool Use Not comparable
Grok-2 (Dec 2024): —, Grok 4 Fast: 29.5 (#86)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| τ²-bench Banking | — | 15.7% |
| Cybench | — | 30% |
| LMArena Search | — | 1171 |
Reasoning Grok 4 Fast leads
Grok-2 (Dec 2024): 16.9 (#299), Grok 4 Fast: 22.2 (#201)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Hard Prompts | 1272 | 1412 |
| DTBench | 65.2% | 82.7% |
| Epoch Capabilities Index | 130.48 | 144.2 |
| ARC-AGI-2 | — | 5.3% |
| SimpleBench | 22.7% | — |
| Kagi LLM Benchmark | — | 66.1% |
| ARC-AGI-1 | — | 48.5% |
| LiveBench Reasoning | 54.8% | — |
| LiveBench Data Analysis | 54.5% | — |
| ForecastBench | — | 60.5 |
| LiveBench | 54.3% | — |
Math Grok 4 Fast leads
Grok-2 (Dec 2024): 20.8 (#284), Grok 4 Fast: 38.9 (#123)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Math | 1283 | 1419 |
| OTIS Mock AIME 2024-2025 | 11.5% | — |
| LiveBench Math | 54.9% | — |
| MATH Level 5 | 63.5% | — |
| FrontierMath (Feb 2025 set) | 0.7% | — |
Knowledge Grok 4 Fast leads
Grok-2 (Dec 2024): 29.8 (#233), Grok 4 Fast: 32.0 (#214)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Expert | 1254 | 1411 |
| GPQA Diamond | 53.8% | — |
| Confabulations | 20.1% | — |
| Vectara Hallucination Rate | — | 19.7% |
Multilingual Grok 4 Fast leads
Grok-2 (Dec 2024): 43.1 (#188), Grok 4 Fast: 51.3 (#111)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Non-English | 1282 | 1396 |
| LMArena Chinese | 1289 | 1457 |
| LMArena French | 1318 | 1431 |
| LMArena German | 1287 | 1383 |
| LMArena Japanese | 1244 | 1352 |
| LMArena Korean | 1237 | 1357 |
| LMArena Russian | 1286 | 1389 |
| LMArena Spanish | 1281 | 1415 |
Instruction Following Grok 4 Fast leads
Grok-2 (Dec 2024): 66.9 (#202), Grok 4 Fast: 73.2 (#121)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Instruction Following | 1270 | 1387 |
| LiveBench Instruction Following | 69.6% | — |
Long Context Grok 4 Fast leads
Grok-2 (Dec 2024): 38.8 (#190), Grok 4 Fast: 63.2 (#3)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Longer Query | 1276 | 1415 |
| Fiction.LiveBench | — | 94.4% |
Writing & Preference Grok 4 Fast leads
Grok-2 (Dec 2024): 48.6 (#198), Grok 4 Fast: 60.0 (#102)
| Benchmark | Grok-2 (Dec 2024) | Grok 4 Fast |
|---|---|---|
| LMArena Text | 1305 | 1407 |
| LMArena Creative Writing | 1284 | 1387 |
| LMArena Multi-Turn | 1290 | 1414 |
| Short-Story Creative Writing | 63.6% | — |
| LiveBench Language | 45.6% | — |
Frequently asked questions
Is Grok-2 (Dec 2024) better than Grok 4 Fast?
Grok 4 Fast is the stronger model overall, scoring 39.4 to 33.7 on the Noometry Index.
Is Grok-2 (Dec 2024) or Grok 4 Fast better for coding?
They score almost the same on coding (33.3 vs 32.5); test both on your own repository before choosing.
How many benchmarks do Grok-2 (Dec 2024) and Grok 4 Fast share?
20 benchmarks have published results for both models. Grok-2 (Dec 2024) has 34 scored results on Noometry and Grok 4 Fast has 30.