vlm.fyi Get the data
Models / Meta

Muse Spark (open)

Measured on 1 provider. The model card lists text, image, video, audio input; what each provider accepts is below.

Fireworks

accounts/fireworks/models/muse-spark-open
Images
120 “Too many images were provided, we currently limit the number of images per conversation to 120”
768×432 frames
112 “Too many images were provided, we currently limit the number of images per conversation to 120”
Tokens / frame
407
Inline body
≥ 32 MB
Images by URL
14 MB “Timed out downloading image <url>, timeout is 10000 ms”
Video file
No video “Muse Spark serving does not support video inputs yet”
Video tok/min
No “Muse Spark serving does not support video inputs yet”
Frames/min
—
Video input
No
Tool messages
Lenient
Extra fields
Rejects usage, reasoning
  • Rejects the OpenRouter-style `usage` and `reasoning` request fields. Full agent sessions with 100 frames got no response within 300 s.

Measured 2026-10-31