Muse Glimmer-30BvsMuse Spark
Across 4 shared benchmarks, Muse Spark leads overall: Muse Glimmer-30B wins 0, Muse Spark wins 4, with 0 ties and an average score difference of -12.52.
Muse Glimmer-30B
Facebook AI研究实验室 · 2026-08-10 · Reasoning model
Muse Spark
Facebook AI研究实验室 · 2026-04-08 · Reasoning model
Muse Glimmer-30B0 wins(0%)(100%)4 winsMuse Spark
Benchmark scores
Grouped by capability, sorted by largest gap within each. 4 shared benchmarks.
AI Agent - Tool Usage
Muse Spark 1/1| Benchmark | Muse Glimmer-30B | Muse Spark | Diff |
|---|---|---|---|
| MCP-Atlas | 75.5020 / 38Thinking High (With Tools) | 82.206 / 38Normal (With Tools) | -6.70 |
Coding and Software Engineer
Muse Spark 1/1| Benchmark | Muse Glimmer-30B | Muse Spark | Diff |
|---|---|---|---|
| SWE-bench Verified | 7637 / 114Thinking High (With Tools) | 77.4027 / 114Thinking (With Tools) | -1.40 |
General Evaluation
Muse Spark 1/1| Benchmark | Muse Glimmer-30B | Muse Spark | Diff |
|---|---|---|---|
| GPQA Diamond | 83.5098 / 226Thinking High (No Tools) | 89.5045 / 226Thinking (No Tools) | -6 |
General Knowledge
Muse Spark 1/1| Benchmark | Muse Glimmer-30B | Muse Spark | Diff |
|---|---|---|---|
| HLE | 22117 / 181Thinking High (No Tools) | 587 / 181深度思考（无工具、并行） | -36 |
Specs
| Field | Muse Glimmer-30B | Muse Spark |
|---|---|---|
| Publisher | Facebook AI研究实验室 | Facebook AI研究实验室 |
| Release date | 2026-08-10 | 2026-04-08 |
| Model type | Reasoning model | Reasoning model |
| Architecture | Dense | Dense |
| Parameters | 29.6B | Not available |
| Context length | 128K+ | 262K |
| Max output | Not available | Not available |
Summary
- Muse Sparkleads in:AI Agent - Tool Usage (1/1), Coding and Software Engineer (1/1), General Evaluation (1/1), General Knowledge (1/1)
On average across the 4 shared benchmarks, Muse Spark scores 12.52 higher.
Largest single-benchmark gap: HLE — Muse Glimmer-30B 22 vs Muse Spark 58 (-36).
Page generated from structured model, pricing and benchmark records. No real-time LLM is used to write the prose.