Compare Models

Muse Spark 1.3 Max

Muse Spark 1.3 Max is a model from Meta, released September 5, 2026. It ranks #9 of 44 models on the Vals Index with 58.16%. Its best result is #1 of 74 on Legal Research Bench.

Release Date: Sep 5, 2026

Developer Meta  🇺🇸
Context Window 1M
Max Output Tokens 131k
Token Costs (in/out) $1.25/4.25
Weights Private
Input Modalities

Accuracy

58.16 % ± 1.19

Cost / Test (Vals Index)

$ 3.787

Latency

23 min 34 s

Fallback Rate 0.00%
Refusal Rate 0.65%
Vals Index
BenchmarksAccuracyRankings
Vals Index

58.16%

±1.19
9/44
Vals RSI Index

19.64%

17/24
Code Migration

47.41%

±4.26
13/74
CUA-bench

5.83%

6/8
CyberBench v1.1

72.74%

±5.67
13/44
EMB

67.43%

±3.06
15/71
Finance Agent (v2)

59.96%

±2.06
4/75
Legal Research Bench

55.29%

±3.46
1/74
MysteryMechanism

36.04%

±3.23
9/24
ProofBench v1.1

58.00%

±4.96
21/47
SRE Bench

3.82%

±1.19
19/32
Tax Agent Bench

72.44%

±2.88
6/66
Vibe Code Bench 1-100

20.46%

±3.85
6/21
Vibe Code Bench v1.1

85.86%

±2.51
11/109
Harvey's Legal Agent Benchmark

23.75%

±3.55
2/75
IOI

56.56%

±2.52
18/41
ProgramBench

2.50%

±1.11
8/62
Terminal-Bench 4.0

24.75%

±0.51
16/44
Terminal-Bench Science

10.00%

±3.61
13/38
Proprietary BenchmarksAcademic BenchmarksIndustry Partners
Vals
Default Provider : Meta
Temperature: 1
Top P: Default
Top K: Default
Max Output Tokens: 131,072
Reasoning Effort: max

Compared with

Muse Spark 1.3 Max vs Claude Fable 5 13 shared benchmarks Muse Spark 1.3 Max vs GPT-5.6 Sol 19 shared benchmarks Muse Spark 1.3 Max vs Claude Opus 5.5 19 shared benchmarks Muse Spark 1.3 Max vs Claude Sonnet 5.5 16 shared benchmarks Muse Spark 1.3 Max vs Claude Fable 5.1 19 shared benchmarks Muse Spark 1.3 Max vs Claude Opus 5 19 shared benchmarks Muse Spark 1.3 Max vs GPT-6 Sol 17 shared benchmarks Muse Spark 1.3 Max vs GPT-6 Astra 19 shared benchmarks Muse Spark 1.3 Max vs Gemini 3.8 Flash 19 shared benchmarks