# Grok 4 Fast (Reasoning) Benchmark Scores & Performance

> Grok 4 Fast (Reasoning) by xAI scores 54.56/100 overall, ranking #104 out of 411 AI models.

Canonical page: https://benchlm.ai/models/grok-4-fast-reasoning

Last updated: September 4, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | xAI |
| Source Type | Proprietary |
| Reasoning Type | Reasoning |
| Context Window | 2M |
| Overall Score | 54.56/100 |
| Overall Rank | #104 of 411 |

## Family & Coverage

- Family: Grok 4 Fast
- Variant: reasoning (reasoning)
- Benchmarks covered: 13 of 422
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Agentic Benchmarks

| Benchmark | Score |
|-----------|-------|
| [τ²-bench results](/benchmarks/tau2-bench) | 65.8% |

## Coding Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Vibe Code Bench](/benchmarks/vibecodebench) | 0.00% |
| [AA-SciCode](/benchmarks/aascicode) | 44.2% |

## Multimodal & Grounded Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-MMMU-Pro](/benchmarks/aammmupro) | 61.8% |

## Reasoning Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-LCR](/benchmarks/lcr) | 71.0% |
| [CritPt](/benchmarks/critpt) | 2.9% |

## Knowledge Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Artificial Analysis Intelligence Index](/benchmarks/artificialanalysis) | 27.9% |
| [AA-GPQA Diamond](/benchmarks/aagpqadiamond) | 84.7% |
| [AA-HLE](/benchmarks/aahle) | 19.1% |
| [AA-Omniscience Index](/benchmarks/aaomniscienceindex) | -29.9% |
| [AA-Omniscience Accuracy](/benchmarks/omniscienceaccuracy) | 22.8% |
| [AA-Omniscience Hallucination Rate](/benchmarks/omnisciencehallucinationrate) | 68.3% |

## Instruction Following Benchmarks

| Benchmark | Score |
|-----------|-------|
| [AA-IFBench](/benchmarks/aaifbench) | 50.5% |

## Other xAI Models

- [Grok 4.6](/models/grok-4-6) - Score: 70.19
- [Grok 4.5](/models/grok-4-5) - Score: 68.64
- [Grok 4.20](/models/grok-4-20-beta) - Score: 67.01
- [Grok 4.3](/models/grok-4-3) - Score: 60.37
- [Grok 4.1](/models/grok-4-1) - Score: 59.33
- [Grok 4.1 Fast (Reasoning)](/models/grok-4-1-fast-reasoning) - Score: 58.18
- [Grok 4](/models/grok-4) - Score: 54.91
- [Grok 3 [Beta]](/models/grok-3-beta) - Score: 51.65
- [Grok 4.1 Fast](/models/grok-4-1-fast) - Score: 49.8
- [Grok 3 Mini](/models/grok-3-mini) - Score: 48.26
- [Grok Code Fast 1](/models/grok-code-fast-1) - Score: 37.21
- [Grok Build 0.1](/models/grok-build-0-1) - Score: not computed
- [Grok TTS](/models/grok-tts) - Score: not computed
- [Grok 4.20 Multi-agent](/models/grok-4-20-multi-agent-beta) - Score: not computed
- [Grok Realtime](/models/grok-realtime) - Score: not computed
- [Grok Voice Think Fast 1.0](/models/grok-voice-think-fast-1-0) - Score: not computed
- [Grok Voice Think Fast 2.0](/models/grok-voice-think-fast-2-0) - Score: not computed