Muse Glimmer 30B VRAM requirements

How much VRAM does Muse Glimmer 30B need?

Muse Glimmer 30B needs 19.2 GB at Q4_K_M, 33.1 GB at Q8_0 and 61.7 GB at FP16 with 8,192 tokens of context and one request. Each total below is the weights plus the FP16 KV cache and the runtime overhead (0.5 GB plus 10%), and each row opens the calculator with that setting. What the GGUF names mean.

Fine-tuning Muse Glimmer 30B? Muse Glimmer 30B VRAM for LoRA, QLoRA and full training.

Muse Glimmer 30B GGUF files on Hugging Face

The 8 GGUF files llama.cpp picks for Muse Glimmer 30B with -hf <repo>:<QUANT>, with the bytes the Hugging Face file list reported on 2026-09-29. Each total is that file plus the FP16 KV cache and 0.5 GB + 10%, one request. "vs estimate" compares the file with the bits-per-weight size the table above uses.

GB here is GiB (1024³ bytes), as everywhere on this site; Hugging Face shows the Q4_K_M file as 17.3 GB, in 10⁹ bytes.

Which Muse Glimmer 30B GGUF fits a 16, 24, 32, 48 or 80 GB GPU or a Mac

From the file sizes above: the largest file that fits each machine with an 8K context and 0.5 GB left free, the longest context it then has room for, and the longest context for Q4_K_M (16.1 GB). One request, FP16 KV cache, -np 1. The Q4_K_M file is 676 MB under its bits-per-weight estimate; the estimate tables and the llama-server command on this page count the estimate, so they show less context.

Run Muse Glimmer 30B with llama.cpp

llama-server -hf bartowski/Muse-Glimmer-30B-GGUF:Q4_K_M -c 131072 -np 1

Muse-Glimmer-30B-Q4_K_M.gguf, 17.3 GB, from bartowski/Muse-Glimmer-30B-GGUF (checked 2026-09-29). On a 24 GB RTX 3090 / 4090, Q4_K_M with -c 131072 takes 20.8 GB and leaves 3.15 GB. -np 1: one slot, so one 2,048-token window.

Muse Glimmer 30B at 8K, 32K and 128K (full) tokens of context

13 of its 52 layers use full attention and 39 keep a sliding window of 2,048 tokens. Each extra token of context adds 13 KB of FP16 cache per request once the sliding windows are full. The last column is the smallest setup here, one card or a group, that holds it at Q4_K_M. How the KV cache works.

How fast Muse Glimmer 30B writes

Tokens per second for one request with 8,192 tokens of context, estimated from memory bandwidth. A dash means it does not fit on one card. Try other settings in the speed calculator.

Longest context on one GPU

How many tokens of context Muse Glimmer 30B fits on a single card with one request, an FP16 KV cache and 0.5 GB left free. "Full" means the model's whole context window fits.

Model details

Parameters

29.8B (29,776,626,688)

Layers

13 of its 52 layers use full attention and 39 keep a sliding window of 2,048 tokens

Attention cache

2 KV heads × 128

Context length

131,072 tokens

Published weights

55.5 GB (BF16)

On Hugging Face

meta-models/Muse-Glimmer-30B

Reported benchmarks

None published in the model card for GPQA, MMLU-Pro or LiveCodeBench.

Muse Glimmer 30B next to similar-sized models

The three models here closest to it in parameter count, at Q4_K_M with 8,192 tokens of context.