Inference
EN

Grok 3 [Beta] vs Grok 3 Mini

Change models
Grok 3 [Beta]xAI
Grok 3 MinixAI
IndexAll categories, out of 100.
39.3 ±8.5
44.3 ±9.7 best
GapIndex points below the top.
−5.0 · no clear gap
Top
CoverageIndex weight that has results.
60%
60%
SpeedOutput tokens per second.
—
65/s
EloLMArena rating and rank.
—
1370
ReleasedMonth of release.
2025-02
2025-02
Input / 1MUS dollars per 1M input tokens.
—
$0.30
Output / 1MUS dollars per 1M output tokens.
—
$0.50
1K chats2K tokens in, 500 out, each.
—
$0.85
1K documents20K tokens in, 1K out, each.
—
$6.50
1K agent tasks200K tokens in, 20K out, each.
—
$70.00
ContextMaximum tokens in one request.
128K
128K
AgenticMulti-step tasks with tools.
—
—
CodingCode writing and repair.
34.9
51.4 best
ReasoningLogic problems and puzzles.
34.4
37.0 best
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
47.1
48.7 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
—
—
MathMath problems.
38.2
43.9 best

Benchmarks

13 shared
Only Grok 3 [Beta]: 13
Grok 3 [Beta]
Grok 3 Mini
GPQA diamond · Epoch AIKnowledge
75.8% best103 / 217
75.5%106 / 217
OTIS Mock AIME 2024-2025Math
55.6%116 / 200
77.8% best90 / 200
Graduate-Level Google-Proof Q&AKnowledge
75.4%79 / 141
80.3% best65 / 141
ARC-AGI-2 (semi-private)Reasoning
0.0%120 / 131
0.4% best117 / 131
ARC-AGI-1 (semi-private)Reasoning
5.5%119 / 126
16.5% best108 / 126
ARC-AGI-2 (public eval)Reasoning
0.0%102 / 122
0.0%102 / 122
ARC-AGI-1 (public eval)Reasoning
8.4%111 / 116
17.6% best103 / 116
MATH level 5Math
88.7% best17 / 88
88.1%18 / 88
FrontierMath-2025-02-28-PrivateMath
3.8%50 / 69
5.9% best43 / 69
FrontierMath-2025-02-28-PublicMath
0.0%33 / 64
0.0%33 / 64
FrontierMath-Tier-4-2025-07-01-PublicMath
0.0%3 / 36
0.0%3 / 36
American Invitational Mathematics Examination 2024Math
52.2%22 / 27
89.5% best8 / 27
LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for CodeCoding
57.0%7 / 12
74.8% best2 / 12