BenchLeader
MetaReasoning modelNewReleased 12 Oct 2026Compare with another model

Muse Wonder

Muse Wonder is a Meta proprietary reasoning model, released 12 Oct 2026. Its best configuration (max reasoning effort) ranks #31 of 761 on the BenchLeader Index at 66.3 ±8.5, in the upper half. The ± is the point: 161 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (88) and lowest in long context (66). At $8.00 per million tokens blended it is among the most expensive ranked models. Output speed of 85 tokens per second puts it faster than most, with a first answer in 60.8 s. Last measured 12 Oct 2026.

Blended price
$8.00/M
$4.00 in · $20.00 out
Output speed
85 tok/s
measured by Artificial Analysis
First answer
61 s
first token 4.62 s
Context
1.0M
Full answer
67 s
median, reasoning included
Cost per run
$4.52
one full Intelligence Index run
Released
12 Oct 2026
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index66
  2. Reasoning88
  3. Knowledge73
  4. Long context66
  5. Composite78

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

BenchmarkmaxSource
LiveBench Reasoningnot in index90.6%#11LiveBench
Humanity's Last Exam (AA)not in index50.1%#24Artificial Analysis
CritPt28.3%#26Artificial Analysis

Coding

BenchmarkmaxSource
LiveBench Codingnot in index80.7%#18LiveBench
SciCode (AA)not in index60.3%#11Artificial Analysis
Terminal-Bench 4.0 (AA)not in index52.0%#15Artificial Analysis

Agents & tools

BenchmarkmaxSource
LiveBench Agentic Codingnot in index62.7%#9LiveBench
GDPval-AA v2.1not in index63.5%#5Artificial Analysis
AutomationBenchnot in index68.1%#7Artificial Analysis
GDP.pdfnot in index30.6%#8Artificial Analysis
AA-Briefcase v1.1not in index1659#9Artificial Analysis

Maths

BenchmarkmaxSource
LiveBench Mathematicsnot in index96.5%#8LiveBench

Instruction following

BenchmarkmaxSource
LiveBench Languagenot in index82.9%#21LiveBench
LiveBench Instruction Followingnot in index73.0%#17LiveBench

Long context

BenchmarkmaxSource
AA-LCR80.7%#62Artificial Analysis

Where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

ProviderSpeedFirst tokenInput $/MOutput $/MContextQuantisation
Meta60 tok/s4.62 s$4.00$20.001.0M–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache. Reasoning tokens are not modelled.

WorkloadTokens in / outCostWith cachingTime
Chat reply400 / 300$0.0076–1.1 min
Summarise a 30-page report12,000 / 600$0.060–1.1 min
Code edit6,000 / 1,500$0.054–1.3 min
Agentic coding session60,000 / 4,000$0.320–1.8 min
Structured extraction2,000 / 200$0.012–1.1 min

More from Meta

New this month

Compare Muse Wonder with…

See also

Data as of 13 Oct 2026. Compare with another model.

Cite as: BenchLeader, “Muse Wonder: benchmarks, pricing, speed and rank”, https://www.benchleader.com/models/muse-wonder, data as of 13 Oct 2026.