← Leaderboard
Muse-Glimmer 30B (GGUF, kquant 17GB)
tested-offloaded Rank #42 of 51 · 4/8 GAUNTLET progress
Hollow markers & dashed spokes: axis not yet scored
G
74
A
76
U
—
N
—
T
100
L
—
E
—
T
5.9
Specification
- Parameters
- 30B (unconfirmed -- filename-derived, see note)
- Architecture
- muse-glimmer
- Size on disk
- 17 GB
- Quantization
- kquant (unconfirmed exact scheme -- filename says "kquant", not a standard llama.cpp quant label)
- Format
- GGUF
- Runtime
- LM Studio
- Class
- Standard
- Reasoning (CoT)
- No
- Internal ID
- M24
- Mean speed
- 20.7 tok/s across suites
- Stall census
- 0 stalls in 22 observed tests (0.0%)
- Model card
- huggingface.co
Suite results
Suite Score Avg / 20 Tests n tok/s Run
General capability (13-task real-workload suite) 1a 192.4 / 260 14.8 13/13 n=1.1 15.4 run 2026-08-13
| Test | Category | Score | tok/s | |
|---|---|---|---|---|
| A1 | Dev/Ops Scripting | 13/20 | 20.6 | |
| A2 | Dev/Ops Scripting | 15/20 | 12.1 | |
| A3 | Dev/Ops Scripting | 15/20 | 14.3 | |
| A4 | Dev/Ops Scripting | 12/20 | 16.1 | |
| A5 | Dev/Ops Scripting | 17/20 | 16.1 | |
| A6 | Dev/Ops Scripting | 12/20 | 16.3 | |
| B1 | Document Processing | 17/20 | 13.4 | |
| B2 | Document Processing | 15/20 | 13.6 | |
| B3 | Document Processing | 20/20 | 14.8 | |
| C1 | Content Production | 15/20 | 16.2 | |
| C2 | Content Production | 16/20 | 16.7 | |
| C3 | Content Production | 14/20 | 14.7 | |
| C4 | Content Production | 11/20 | 15.5 |
Category mean Dev/Ops Scripting 14Document Processing 17.3Content Production 14
Agentic tool-calling & protocol adherence 1b 121.6 / 160 15.2 8/8 n=1 26 run 2026-08-13
| Test | Category | Score | tok/s | |
|---|---|---|---|---|
| WA1 | Tool Calling & Protocol Adherence | 11/20 | 26 | |
| WA2 | Tool Calling & Protocol Adherence | 14/20 | 21.2 | |
| WA3 | Tool Calling & Protocol Adherence | 14/20 | 24.2 | |
| WA4 | Tool Calling & Protocol Adherence | 17/20 | 14.4 | |
| WA5 | Tool Calling & Protocol Adherence | 17/20 | 24.5 | |
| WB1 | Agent Robustness & State Management | 16/20 | 22.4 | |
| WB2 | Agent Robustness & State Management | 18/20 | 18.1 | |
| WB3 | Agent Robustness & State Management | 15/20 | 16.9 |
Category mean Tool Calling & Protocol Adherence 14.6Agent Robustness & State Management 16.3
Each test is LLM-judged 0–20 against a fixed rubric; suite max = tests × 20. An ADJ chip marks a cell carrying a subjective adjustment from the maintainer; the judged score is the one shown, is unchanged, and remains what the axes, ranks and tie bands are computed from. Hover for the reason. Rows with a ▶ expand to the per-test breakdown. Where fewer tests ran than the suite total, unrun tests count as zero toward the axis score. See methodology.