- Parameters
- 30B
- Quant
- UD-Q4_K_XL
- Size
- 14.8 GB
- Family
- muse-glimmer
muse-glimmer-30b
Measured, not seateddenseresidentllama.cpp
Model card ↗Outcome as of 20 September 2026, from the seat registry. The write-up below was made when the model was tested.
Why it was tested
Previously dismissed at 10.4 tok/s without ever being scored. Re-examined because it is multimodal, and scanned invoices are the part of the accountancy job nothing else can read.
What came of it
Reproduced at 9.6 tok/s, so the speed verdict stands. But its 32:2 head ratio makes it the most context-efficient model here: 10.6GB of KV at full context against granite's 32GB. Its nullschema is unmeasurable because llama.cpp silently drops the schema for this model specifically. Confirmed on the ladder as the cheapest context here: 1.77 GB of KV across the full 131k window, against granite's 30.26 GB.
| Suite | Score | 95% interval | Passed | Runtime · profile | Recorded |
|---|---|---|---|---|---|
| grounding | 85.6% | 78.4–90.7 | Passed 107/125 | llama.cppextraction | 2026-09-08 |
| screening | 86.7% | 75.8–93.1 | Passed 52/60 | llama.cppextraction | 2026-09-08 |
Runs marked unenforced were measured through mlx_lm on a suite that requires constrained decoding. mlx_lm accepts the schema and ignores it, so those numbers do not measure what the suite is designed to measure. They are kept and labelled rather than deleted.