xAI · model passport
Grok 4 Fast
grok-4-fast- Cloud only
- Proprietary licence
- Released 2025-09-19(curated)
- Best percentile
- 93.5LiveCodeBench (Artificial Analysis)
- Cohort standing
- 53.9median 52.2
What the data says
Benchmarked on 5 of 27 tracked axes. Every figure below is the model's own published result on that board.
Strongest axes
- 93.5 pct — LiveCodeBench (Artificial Analysis) (83.2%)
- 91.4 pct — MMLU-Pro (Artificial Analysis) (85.0%)
Weakest axes
- 2.9 pct — WebDev Arena (1,160)
Uneven: 90.6 percentile points between its best and worst axis — 93.5 pct on LiveCodeBench (Artificial Analysis) versus 2.9 pct on WebDev Arena.
- LiveCodeBench (Artificial Analysis)93.5 pct83.2%Verify →
- MMLU-Pro (Artificial Analysis)91.4 pct85.0%Verify →
- IFBench (Artificial Analysis)57.7 pct50.5%Verify →
- ARC-AGI-236 pct5.3%Verify →
- WebDev Arena2.9 pct1,160Verify →
- OverallNot benchmarked
- Vision ArenaNot benchmarked
- Factuality ArenaNot benchmarked
- Search ArenaNot benchmarked
- Document ArenaNot benchmarked
- Agent ArenaNot benchmarked
- GPQA DiamondNot benchmarked
- AIME (math)Not benchmarked
- FrontierMathNot benchmarked
- SimpleQA VerifiedNot benchmarked
- Humanity's Last ExamNot benchmarked
- SimpleBenchNot benchmarked
- LiveBenchNot benchmarked
- BFCL v4 (function calling)Not benchmarked
- SWE-bench Verified (live board)Not benchmarked
- Terminal-Bench 4.0Not benchmarked
- HHEM Hallucination (Vectara)Not benchmarked
- SWE-bench VerifiedNot benchmarked
- Terminal-BenchNot benchmarked
- CodingNot benchmarked
- Instruction followingNot benchmarked
- Hard promptsNot benchmarked
Percentiles are the model's standing within each board's own population. Axes marked "not benchmarked" have no published result for this model — that is an absence of data, not a low score.
Grok 4 Fast (Reasoning) stands at 53.9 (raw mean 56.3) — measured on 5 of 27 axes.
Grok 4 Fast (Reasoning) stands at 53.9 versus a cohort median of 52.2 across 147 models released within 180 days — above the peer median.
Grok 4 Fast (Reasoning) scores at percentile 93.5 on LiveCodeBench (Artificial Analysis).
Grok 4 Fast (Reasoning) scores at percentile 91.4 on MMLU-Pro (Artificial Analysis).
Grok 4 Fast (Reasoning) scores at percentile 2.9 on WebDev Arena.
Grok 4 Fast (Reasoning) shows a 90.6-point percentile gap across 5 covered axes: 93.5 on LiveCodeBench (Artificial Analysis) versus 2.9 on WebDev Arena.
Every claim is a computed fact from the current board data, phrased by a low-temperature model call and rejected by a validation gate if any number is not in the source row. How the leaderboards work →