Muse-Glimmer-30B-ctx96k-kvq8 on RTX A4000 (16GB)
MeasuredOn 2026-10-22, Muse-Glimmer-30B-ctx96k-kvq8 UD-IQ3_XXS ran at 21.8 tok/s (chat-short) on an RTX A4000 (16GB) with llama.cpp. Run r_0o1xfrrn-ch.
Workload results
| Workload | Backend | Model | decode tok/s | prefill tok/s | TTFT | p50 | p95 |
|---|---|---|---|---|---|---|---|
| chat-short | llama.cpp@0.7.0-dev (build 11803, commit a37cc4610) | Muse-Glimmer-30B-ctx96k-kvq8UD-IQ3_XXS | 21.78tok/s | 264.1tok/s | 583ms | 46.0ms | 47.2ms |
| chat-long | llama.cpp@0.7.0-dev (build 11803, commit a37cc4610) | Muse-Glimmer-30B-ctx96k-kvq8UD-IQ3_XXS | 20.12tok/s | 695.0tok/s | 4,517ms | 49.4ms | 50.8ms |
| concurrent-decode | llama.cpp@0.7.0-dev (build 11803, commit a37cc4610) | Muse-Glimmer-30B-ctx96k-kvq8UD-IQ3_XXS | 20.90tok/s | — | — | 47.8ms | 48.9ms |
| agent-trace | llama.cpp@0.7.0-dev (build 11803, commit a37cc4610) | Muse-Glimmer-30B-ctx96k-kvq8UD-IQ3_XXS | 20.18tok/s | 1,626.9tok/s | 1,013ms | 49.2ms | 52.4ms |
Reproduce on your machine
Start with this model and workload selection. This command does not capture the original artifact, backend version or server settings:
Match quantization, context capacity, caching and runtime settings for a useful comparison. Duration varies by workload and hardware; the published client may differ from this run's version. How it's measured.
Share this run
Copy the short link, a Reddit-ready summary, or a README badge. Every number is the measured result on this page, and each badge links back to it.
https://llm-speed.com/r/r_0o1xfrrn-chRelated benchmarks
- More Muse-Glimmer-30B-ctx96k-kvq8 benchmarks — every backend and rig that has run this model.
- More RTX A4000 (16GB) LLM benchmarks — every model measured on this hardware.
Provenance
- Run ID
- r_0o1xfrrn-ch
- Fingerprint hash
- Public key
- 7OUg4nnOCjh+MfVQqxaCH+zYo4Rj3yJARN4rngyKmwc=
- Received
- 2026-10-22 14:04:31