Hollow markers & dashed spokes: axis not yet scored
G
89
A
95
U
—
N
—
T
100
L
—
E
71.5
T
41.2
Specification
- Parameters
- 80B
- Architecture
- qwen3_next
- Size on disk
- 44.86 GB
- Quantization
- 4bit
- Format
- MLX
- Reasoning (CoT)
- No
- Internal ID
- M6
- Mean speed
- 67.3 tok/s across suites
- Stall census
- 0 stalls in 46 observed tests (0.0%)
- Model card
- lmstudio.ai
Suite results
| Suite | Score | Avg / 20 | Tests | tok/s |
|---|---|---|---|---|
| General capability (13-task real-workload suite) 1a | 231.4 / 260 | 17.8 | 13/13 | 77 |
| Agentic tool-calling & protocol adherence 1b | 152 / 160 | 19 | 8/8 | 46.1 |
| Coding depth 1c | 85.8 / 120 | 14.3 | 6/6 | 69.3 |
| Content-production depth 1e | 109.2 / 120 | 18.2 | 6/6 | 76.9 |
Each test is LLM-judged 0–20 against a fixed rubric; suite max = tests × 20. Where fewer tests ran than the suite total, unrun tests count as zero toward the axis score. See methodology.