MiniMax's fast multimodal agent and coding model with tunable thinking depth, the generally available successor of the M3.1 Flash Preview. it costs half of MiniMax M3's sub-512K per-token rate.
pick it if
you want MiniMax at half of M3's sub-512K per-token rate
you send prompts past 512K tokens; the rate stays flat
your agent re-reads long context; cached input bills at a tenth
skip it if
you need JSON-schema constrained output
you need downloadable weights today; MiniMax says they come later
MiniMax M3 is MiniMax's open-weight model (about 428B total, 23B active parameters) for coding and agent work, with image and video input. in this catalog it costs twice MiniMax M3.1 Flash, the newer October 2026 release.
pick it if
you need image and video input at $0.30 / $1.20 per million
you want open weights you can also self-host
you reuse long prompts: cache hits bill at a fifth of input
skip it if
MiniMax M3.1 Flash, at half the price, covers your workload
you need JSON-schema constrained output
faq
which is cheaper, MiniMax M3.1 Flash or MiniMax M3?
MiniMax M3.1 Flash lists lower input and output prices. on a coding-agent session (2M tokens in, 60,000 out) MiniMax M3.1 Flash costs $0.34 and MiniMax M3 $0.67.
can i switch between MiniMax M3.1 Flash and MiniMax M3 mid-task?
yes. in an octomind session, /model octohub:minimax-m3.1-flash and /model octohub:minimax-m3 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.