MiniMax offers M3.1 Flash through hosted tools, with weights still pending

MiniMax’s M3.1 Flash is available through its API, M Plan and MiniMax Code, while a weights release is still planned. The company lists pricing of $0.15 per million input tokens, $0.60 for output and $0.015 for cached input. Its coding and tool-use claims come from the vendor, not an independent evaluation.

MiniMax · Oct 19, 2026

Why it matters

A low token price is useful to investigate only alongside task success, retries and total usage.

More context

This October 19 release is an older discovery. Try a small hosted evaluation with representative coding tasks and a fixed spending limit. Compare correct completed work and latency against your existing model. Self-hosting must wait for actual weights and their deployment requirements; the announcement alone does not provide that option.

Keep in mind

Vendor benchmark results do not establish performance in your repository. Token prices alone omit retries, tool overhead and the cost of reviewing incorrect work.

Original read