Meta Superintelligence Labs · meta/muse-glimmer-30b
| Provider | Meta Superintelligence Labs |
|---|---|
| Family | muse |
| Type | vlm |
| Status | active |
| Released | 2026-08-10 |
| Updated | 2026-08-11 |
| Parameters | 29.8B |
| Open weights | yes |
| Relationship | Model |
|---|---|
| Is a distillation of | Muse Spark |
Chain Of Thought · tier-2 Function Calling · tier-2 Multi Step · tier-2 Multilingual · tier-2 Think Budget Control · tier-2 Tool Error Recovery · tier-2
Ordered by predicted speed. Bandwidth sets decode rate; memory decides whether it runs at all.
| Device | Memory | Bandwidth | Best quality | Weights | tok/s | Fastest |
|---|---|---|---|---|---|---|
| Cerebras WSE-3 | 44.0 GB | 21000000.0 GB/s | fp8 | 29.78 GB | ~493675.8 | ~987351.6 q4 |
| NVIDIA GB200 Grace Blackwell Superchip | 372.0 GB | 16000.0 GB/s | bf16 | 59.55 GB | ~188.1 | ~752.3 q4 |
| AMD Instinct MI355X | 288.0 GB | 8000.0 GB/s | bf16 | 59.55 GB | ~94.0 | ~376.1 q4 |
| NVIDIA B200 | 180.0 GB | 7700.0 GB/s | bf16 | 59.55 GB | ~90.5 | ~362.0 q4 |
| Google TPU7x (Ironwood) | 192.0 GB | 7380.0 GB/s | bf16 | 59.55 GB | ~86.7 | ~347.0 q4 |
| AMD Instinct MI325X | 256.0 GB | 6000.0 GB/s | bf16 | 59.55 GB | ~70.5 | ~282.1 q4 |
| AMD Instinct MI300X | 192.0 GB | 5300.0 GB/s | bf16 | 59.55 GB | ~62.3 | ~249.2 q4 |
| NVIDIA H200 SXM | 141.0 GB | 4800.0 GB/s | bf16 | 59.55 GB | ~56.4 | ~225.7 q4 |
| NVIDIA H100 NVL | 94.0 GB | 3900.0 GB/s | bf16 | 59.55 GB | ~45.8 | ~183.4 q4 |
| NVIDIA H100 SXM | 80.0 GB | 3350.0 GB/s | bf16 | 59.55 GB | ~39.4 | ~157.5 q4 |
| AMD Instinct MI250X | 128.0 GB | 3200.0 GB/s | bf16 | 59.55 GB | ~37.6 | ~150.5 q4 |
| Google TPU v5p | 95.0 GB | 2765.0 GB/s | bf16 | 59.55 GB | ~32.5 | ~130.0 q4 |
| NVIDIA A100 80GB SXM | 80.0 GB | 2039.0 GB/s | bf16 | 59.55 GB | ~24.0 | ~95.9 q4 |
| NVIDIA H100 PCIe | 80.0 GB | 2000.0 GB/s | bf16 | 59.55 GB | ~23.5 | ~94.0 q4 |
| NVIDIA GeForce RTX 5090 | 32.0 GB | 1792.0 GB/s | q6 | 22.33 GB | ~56.2 | ~84.3 q4 |
| Google TPU v6e (Trillium) | 32.0 GB | 1638.0 GB/s | q6 | 22.33 GB | ~51.3 | ~77.0 q4 |
| AMD Instinct MI210 | 64.0 GB | 1600.0 GB/s | fp8 | 29.78 GB | ~37.6 | ~75.2 q4 |
| NVIDIA A100 40GB SXM | 40.0 GB | 1555.0 GB/s | fp8 | 29.78 GB | ~36.6 | ~73.1 q4 |
| Google TPU v4 | 32.0 GB | 1200.0 GB/s | q6 | 22.33 GB | ~37.6 | ~56.4 q4 |
| NVIDIA GeForce RTX 4090 | 24.0 GB | 1008.0 GB/s | q4 | 14.89 GB | ~47.4 | ~47.4 q4 |
| AMD Radeon RX 7900 XTX | 24.0 GB | 960.0 GB/s | q4 | 14.89 GB | ~45.1 | ~45.1 q4 |
| NVIDIA GeForce RTX 3090 | 24.0 GB | 936.0 GB/s | q4 | 14.89 GB | ~44.0 | ~44.0 q4 |
| NVIDIA L40S | 48.0 GB | 864.0 GB/s | fp8 | 29.78 GB | ~20.3 | ~40.6 q4 |
| Apple M3 Ultra | 512.0 GB | 819.0 GB/s | bf16 | 59.55 GB | ~9.6 | ~38.5 q4 |
| AMD Radeon RX 7900 XT | 20.0 GB | 800.0 GB/s | q4 | 14.89 GB | ~37.6 | ~37.6 q4 |
| Apple M2 Ultra | 192.0 GB | 800.0 GB/s | bf16 | 59.55 GB | ~9.4 | ~37.6 q4 |
| Apple M4 Max | 128.0 GB | 546.0 GB/s | bf16 | 59.55 GB | ~6.4 | ~25.7 q4 |
| Apple M1 Max | 64.0 GB | 400.0 GB/s | fp8 | 29.78 GB | ~9.4 | ~18.8 q4 |
| Apple M2 Max | 96.0 GB | 400.0 GB/s | bf16 | 59.55 GB | ~4.7 | ~18.8 q4 |
| Apple M3 Max | 128.0 GB | 400.0 GB/s | bf16 | 59.55 GB | ~4.7 | ~18.8 q4 |
| NVIDIA L4 | 24.0 GB | 300.0 GB/s | q4 | 14.89 GB | ~14.1 | ~14.1 q4 |
| Apple M4 Pro | 64.0 GB | 273.0 GB/s | fp8 | 29.78 GB | ~6.4 | ~12.8 q4 |
| Apple M4 | 32.0 GB | 120.0 GB/s | q6 | 22.33 GB | ~3.8 | ~5.6 q4 |
Taken from the card's availability section.
| Platform | Id on that platform |
|---|---|
| Fireworks AI | accounts/fireworks/models/muse-glimmer-30b |
| Hugging Face | meta-models/Muse-Glimmer-30B |
| MLX Community | mlx-community/Muse-Glimmer-30B-OptiQ-4bit |
| OpenRouter | meta/muse-glimmer-30b |
| Together AI | meta-models/Muse-Glimmer-30B |
| Benchmark | Evaluated as | Score | Evidence date | Source kind | |
|---|---|---|---|---|---|
| arena_elo_style_control | muse-glimmer | 1427.34 elo | 2026-09-10 evaluated | independent evaluator | source |
This card reports no benchmark scores yet.
| Benchmark | Catalogue standing | Score |
|---|