Hollow markers & dashed spokes: axis not yet scored
G
94
A
95.5
U
—
N
—
T
100
L
—
E
—
T
7.8
Specification
- Parameters
- 27B dense
- Architecture
- qwen3_5
- Size on disk
- 29.98 GB
- Quantization
- Q8_0
- Format
- GGUF
- Runtime
- LM Studio
- Class
- Standard
- Reasoning (CoT)
- Yes — emits reasoning tokens
- Internal ID
- M25c
- Mean speed
- 27.6 tok/s across suites
- Stall census
- 0 stalls in 21 observed tests (0.0%)
- Reasoning appetite
- 1,614 tokens mean · 11,418 max
- Model card
- huggingface.co
Suite results
Suite Score Avg / 20 Tests n tok/s Run
General capability (13-task real-workload suite) 1a 244.4 / 260 18.8 13/13 n=1 24 run 2026-08-15
| Test | Category | Score | tok/s | |
|---|---|---|---|---|
| A1 | Dev/Ops Scripting | 20/20 | 24.7 | |
| A2 | Dev/Ops Scripting | 20/20 | 22.8 | |
| A3 | Dev/Ops Scripting | 10/20 | 21.5 | |
| A4 | Dev/Ops Scripting | 19/20 | 22.6 | |
| A5 | Dev/Ops Scripting | 20/20 | 23.8 | |
| A6 | Dev/Ops Scripting | 20/20 | 23.8 | |
| B1 | Document Processing | 20/20 | 26.8 | |
| B2 | Document Processing | 20/20 | 28.3 | |
| B3 | Document Processing | 20/20 | 23.8 | |
| C1 | Content Production | 20/20 | 20.8 | |
| C2 | Content Production | 18/20 | 21.4 | |
| C3 | Content Production | 18/20 | 20 | |
| C4 | Content Production | 20/20 | 21.7 |
Category mean Dev/Ops Scripting 18.2Document Processing 20Content Production 19
Agentic tool-calling & protocol adherence 1b 152.8 / 160 19.1 8/8 n=1 31.1 run 2026-08-15
| Test | Category | Score | tok/s | |
|---|---|---|---|---|
| WA1 | Tool Calling & Protocol Adherence | 20/20 | 24.7 | |
| WA2 | Tool Calling & Protocol Adherence | 20/20 | 21.9 | |
| WA3 | Tool Calling & Protocol Adherence | 20/20 | 24.5 | |
| WA4 | Tool Calling & Protocol Adherence | 20/20 | 16.9 | |
| WA5 | Tool Calling & Protocol Adherence | 20/20 | 26.3 | |
| WB1 | Agent Robustness & State Management | 20/20 | 23.7 | |
| WB2 | Agent Robustness & State Management | 20/20 | 23.1 | |
| WB3 | Agent Robustness & State Management | 13/20 | 20.3 |
Category mean Tool Calling & Protocol Adherence 20Agent Robustness & State Management 17.7
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.