xAI
Grok 4 Fast (Reasoning)
Grok 4 Fast (Reasoning) is an xAI Grok model profile, part of xAI's assistant and reasoning model family for real-time, coding, and general-purpose AI workflows. The benchmark snapshot highlights how this Grok variant compares with competing frontier and open-weight models on capability, speed, latency, and cost.
xAI model releasesOperational Metrics
Output Speed76.2 tok/s
First Token8.28s
Blended Price$0.275/M
Model Metadata
Queryable facts extracted from the upstream model payload.
ReleaseSep 19, 2025
Context Windown/a
Modalitiesn/a
Strength: LCB
Rank #20 across 343 models.
83.2%
Strength: MMLU-Pro
Rank #33 across 345 models.
85.0%
Strength: SciCode
Rank #45 across 472 models.
44.2%
Watch Area: TTFT
Rank #252 across 293 models.
8.28s
Strength Profile
Percentile score by analysis domain.
* Cost is inverted: lower input, output, and blended prices rank higher.
Benchmark Percentiles
Higher bars mean stronger relative placement.
All Benchmarks
| Metric | Domain | Value | Rank |
|---|---|---|---|
| Artificial Analysis Intelligence Index | overall | 35.1 | #106 |
| Artificial Analysis Coding Index | coding | 27.4 | #130 |
| Artificial Analysis Math Index | math | 89.7 | #28 |
| MMLU-Pro | reasoning | 85.0% | #33 |