Muse Glimmer 30B
Muse Glimmer 30B is a dense, open-weight multimodal model from Meta Superintelligence Labs, distilled from Muse Spark and optimized for autonomous agents on consumer hardware. It is suited for long-horizon...
Leaderboards
Average Score combining domain-specific Autobench scores; Higher is better
Performance vs. Industry Average
Intelligence
Muse Glimmer 30B is of lower intelligence compared to average (4.5), with an intelligence score of 4.4.
Price
Muse Glimmer 30B is cheaper compared to average ($1.13 per 1M Tokens) with a price of $0.11 per 1M Tokens.
Latency
Muse Glimmer 30B has a lower average latency compared to average (47.46s), with an average latency of 17.02s.
P99 Latency
Muse Glimmer 30B has a lower P99 latency compared to average (288.94s), taking 82.39s to receive the first token at P99 (TTFT).
Context Window
Muse Glimmer 30B has a smaller context window than average (453k tokens), with a context window of 131k tokens.