Qwen4 Max is Alibaba's flagship Qwen4 model: a hosted 2.4T-parameter MoE that takes text, image and video, with a thinking mode. it succeeds Qwen3.8 Max at lower rates, and its weights are not published.
pick it if
you want Qwen4's flagship with image and video input
long agent sessions reuse a prefix; cache hits bill $0.075 per 1M, against $0.25 on Qwen3.8 Max
token-heavy agent runs; Alibaba reports 48 to 49% fewer tokens than Qwen3.8 Max on Terminal-Bench 4.0
skip it if
you need downloadable weights; Alibaba has announced open weights for smaller Qwen4 models, none for Max
you need audio input; Qwen3.8 Omni Flash takes audio, Max does not
Muse Spark 1.3 is the latest Muse Spark version on Meta's Model API, tuned for agentic workflows with improved coding over 1.2 at the same rate card. Muse Wonder 1.0, which Meta calls its most capable model, costs more.
pick it if
you want a 1M-context multimodal model well below Muse Wonder's price
you reuse long prompts: cache hits bill at about an eighth of input
you run multi-step agentic tool work, which Meta tuned 1.3 for
skip it if
you accept Meta training on your prompts; the Contributor tier is far cheaper
you need reliable audio understanding; Meta says 1.3 does not fully support it
faq
which is better, Qwen4 Max or Muse Spark 1.3?
on DeepSWE v1.1, as each maker reports it, Qwen4 Max scores 75.7 and Muse Spark 1.3 75.4. these are each maker's own numbers, not a controlled head-to-head; the spec table lists every benchmark both publish.
which is cheaper, Qwen4 Max or Muse Spark 1.3?
Muse Spark 1.3 lists lower input and output prices. on a coding-agent session (2M tokens in, 60,000 out) Qwen4 Max costs $3.27 and Muse Spark 1.3 $2.76.
can i switch between Qwen4 Max and Muse Spark 1.3 mid-task?
yes. in an octomind session, /model octohub:qwen4-max and /model octohub:muse-spark-1.3 switch models without losing the conversation, so you can draft on one and review on the other.
try either model
switch between models inside the same octomind session.