Inference
EN

Grok 4.7 vs Muse Wonder

Change models
Grok 4.7xAI
Muse WonderMeta
IndexAll categories, out of 100.
69.1 ±3.7
73.1 ±4.0 best
GapIndex points below the top.
−4.0 · no clear gap
Top
CoverageIndex weight that has results.
85%
85%
SpeedOutput tokens per second.
51/s
80/s best
EloLMArena rating and rank.
1398
1503 best
ReleasedMonth of release.
2026-09
2026-10
Input / 1MUS dollars per 1M input tokens.
$2.00 best
$4.00
Output / 1MUS dollars per 1M output tokens.
$6.00 best
$20.00
Batch output / 1MBatch price per 1M output tokens.
$4.80
—
1K chats2K tokens in, 500 out, each.
$7.00 best
$18.00
1K documents20K tokens in, 1K out, each.
$46.00 best
$100.00
1K agent tasks200K tokens in, 20K out, each.
$520.00 best
$1,200.00
ContextMaximum tokens in one request.
500K
1.05M best
AgenticMulti-step tasks with tools.
71.9
73.0 best
CodingCode writing and repair.
69.9
72.3 best
ReasoningLogic problems and puzzles.
64.5
73.1 best
MultimodalTasks with images and text.
—
—
KnowledgeFacts and expert knowledge.
66.5
68.6 best
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
78.9 best
74.4
MathMath problems.
61.8
73.5 best

Benchmarks

53 shared
Only Grok 4.7: 31Only Muse Wonder: 9
Grok 4.7
Muse Wonder
Artificial Analysis Humanity's Last ExamKnowledge
43.1%32 / 222
50.1% best12 / 222
Artificial Analysis Long Context ReasoningReasoning
76.7%84 / 217
80.7% best37 / 217
Critical Physics TasksReasoning
17.7%37 / 214
28.3% best15 / 214
OTIS Mock AIME 2024-2025Math
98.1%31 / 200
99.4% best17 / 200
Artificial Analysis Omniscience AccuracyKnowledge
42.7% best42 / 164
38.0%59 / 164
GDPval-AA normalizedAgentic
60.7%8 / 133
63.5% best5 / 133
Vibe Code Bench v1.1Coding
86.2% best13 / 123
83.1%22 / 123
Artificial Analysis SciCodeCoding
57.4%19 / 114
60.3% best10 / 114
FrontierMath-Tiers-1-3-v2-PrivateMath
53.0%53 / 90
82.1% best18 / 90
Legal Research BenchLegal
47.1%16 / 90
50.5% best7 / 90
Tax Agent BenchFinance
35.8%16 / 82
36.3% best15 / 82
LiveBench Data AnalysisReasoning
76.9%30 / 80
81.6% best5 / 80
LiveBench ReasoningReasoning
82.7%51 / 80
90.6% best11 / 80
Mystery Game PuzzlesReasoning
29.0%32 / 80
51.0% best15 / 80
Finance Agent (v2)Finance
52.3%36 / 79
56.6% best18 / 79
Code MigrationCoding
44.8%19 / 77
62.6% best6 / 77
EMBFinance
67.0%20 / 75
67.1% best18 / 75
FrontierMath-Tier-4-v2-PrivateMath
17.1%52 / 72
70.7% best17 / 72
LiveBench Agentic CodingAgentic
54.0%33 / 66
62.7% best9 / 66
LiveBench CodingCoding
77.2%37 / 66
80.7% best16 / 66