Inference
EN

Muse Wonder vs Step 5 Preview

Change models
Muse WonderMeta
Step 5 PreviewStepFun
IndexAll categories, out of 100.
73.1 ±4.0 best
68.4 ±3.7
GapIndex points below the top.
Top
−4.7 · no clear gap
CoverageIndex weight that has results.
85%
90% best
SpeedOutput tokens per second.
80/s
105/s best
EloLMArena rating and rank.
1503 best
1443
ReleasedMonth of release.
2026-10
2026-09
Input / 1MUS dollars per 1M input tokens.
$4.00
$1.00 best
Output / 1MUS dollars per 1M output tokens.
$20.00
$2.70 best
1K chats2K tokens in, 500 out, each.
$18.00
$3.35 best
1K documents20K tokens in, 1K out, each.
$100.00
$22.70 best
1K agent tasks200K tokens in, 20K out, each.
$1,200.00
$254.00 best
ContextMaximum tokens in one request.
1.05M best
1M
AgenticMulti-step tasks with tools.
73.0 best
68.8
CodingCode writing and repair.
72.3 best
71.4
ReasoningLogic problems and puzzles.
73.1 best
67.4
MultimodalTasks with images and text.
—
62.2
KnowledgeFacts and expert knowledge.
68.6 best
63.5
MultilingualTasks in many languages.
—
—
InstructionTasks with strict prompt rules.
74.4
—
MathMath problems.
73.5 best
61.7

Benchmarks

34 shared
Only Muse Wonder: 28Only Step 5 Preview: 28
Muse Wonder
Step 5 Preview
Artificial Analysis Humanity's Last ExamKnowledge
50.1% best12 / 222
46.5%23 / 222
Artificial Analysis Long Context ReasoningReasoning
80.7%37 / 217
88.3% best2 / 217
Critical Physics TasksReasoning
28.3% best15 / 214
20.9%26 / 214
Artificial Analysis Omniscience AccuracyKnowledge
38.0% best59 / 164
36.4%64 / 164
GDPval-AA normalizedAgentic
63.5% best5 / 133
54.3%24 / 133
Vibe Code Bench v1.1Coding
83.1% best22 / 123
70.4%48 / 123
Artificial Analysis SciCodeCoding
60.3% best10 / 114
58.9%13 / 114
Legal Research BenchLegal
50.5% best7 / 90
40.4%35 / 90
Tax Agent BenchFinance
36.3% best15 / 82
27.3%39 / 82
LiveBench Data AnalysisReasoning
81.6% best5 / 80
63.5%69 / 80
LiveBench ReasoningReasoning
90.6% best11 / 80
83.8%44 / 80
Finance Agent (v2)Finance
56.6% best18 / 79
50.7%42 / 79
Code MigrationCoding
62.6% best6 / 77
39.0%31 / 77
EMBFinance
67.1% best18 / 75
57.1%41 / 75
Terminal-Bench 4.0Agentic
49.5% best11 / 61
31.8%23 / 61
DeepSWEAgentic
78.8% best2 / 59
67.7%27 / 59
Agent Arena steerabilityAgentic
2.6 best22 / 54
−0.930 / 54
Agent Arena task outcomeAgentic
5.4 best13 / 54
−1.535 / 54
ProofBench v1.1Math
84.0% best15 / 54
42.0%37 / 54
LiveBench Coding 2.0Coding
86.2% best20 / 52
85.6%22 / 52