Grok 4.1 Fast is xAI's agentic tool-calling model, optimized for real-world use cases such as customer support and deep research. It features a 2M-token context window — the largest among Western frontier models — and focuses on significantly reduced hallucination rates for information-seeking tasks. Reasoning can be enabled or disabled via the API's reasoning parameter, allowing developers to choose between speed-optimized direct answers and deeper analytical responses.
Grok SuperGrokGrok SuperGrok HeavyAPI|VisionReasoningWeb SearchFile|Proprietary Model
Knowledge Cutoff
2024-11
Input → Output Format
Context Memory
2MIN30KOUT
AI Performance Evaluation
Arena Overall Score
1432
±4As of 2026-04-23
Overall Rank
No.57
46,539 Votes
Arena by Ability
Hard Prompts
1442±5No.69
Expert Knowledge
1438±11No.74
Instruction Following
1401±6No.86
Conversation Memory
1416±7No.79
Creative
1412±8No.45
Coding
1464±6No.72
Math
1421±11No.67
Arena by Occupation
Creative Writing
1405±6No.65
Social Sciences
1450±7No.55
Media
1400±7No.54
Business
1417±7No.68
Healthcare
1447±11No.66
Legal
1426±11No.73
Software
1460±5No.62
Mathematics
1422±13No.71
Source:Arena Intelligence
Overall
AA Intelligence Index
24%↓15%
LiveBench
32%↓29%
ForecastBench
54%↓5%
Reasoning & Math
AA Math Index
34%↓39%
GPQA Diamond
64%↓17%
HLE
5.0%↓12%
MMLU-Pro
74%↓8%
AIME 2025
34%↓39%
LB Reasoning
23%↓36%
LB Math
39%↓35%
LB Data
41%↓9%
Coding
AA Coding Index
20%↓15%
LiveCodeBench
40%↓25%
LB Coding
54%↓19%
LB Agentic
10%↓33%
TAU2
64%↓10%
TerminalBench
14%↓17%
SciCode
30%↓11%
Language & Instructions
IFBench
37%↓20%
AA-LCR
22%↓40%
Hallucination (HHEM)
18%↑8%
Factual (HHEM)
82%↓8%
LB Language
50%↓22%
LB IF
17%↓29%
Output Speed
Standard Mode
134tok/s↑52
First Output 0.36s
Reasoning Mode
141tok/s↑53
First Output 11.23s
