OpenCode Console currently lists MiniMax M3.1 Flash Free, under the model ID minimax-m3.1-flash-free, as free for a limited time.
On the official pricing table, input, output, and cached read are Free. Cached write is marked "-", not Free. The official models API gives a 1,000,000-token context window and a 512,000-token output limit. The documentation only says the listing is free for a limited time, and it does not name an end date.
| Model | Input | Output | Cached Read | Cached Write |
|---|---|---|---|---|
| MiniMax M3.1 Flash Free | Free | Free | Free | - |
The provider follows a zero-retention policy and does not use the data for model training. Free-tier chat models have to be used inside OpenCode.
This is not the billed per-token API on the official MiniMax site, where MiniMax-M3.1-Flash is $0.15 input and $0.60 output per million tokens, with cached reads at $0.015. It is also not the paid OpenCode Go model minimax-m3.1-flash, which uses the same unit prices.
The earlier note on MiniMax M3.1 Flash covers the model launch and MiniMax’s official per-token prices, and it recorded a previous three-day free window. The listing here is a different OpenCode Console ID, minimax-m3.1-flash-free, that is free for a limited time — not the same listing or the same price table.
Terms are in the Console Models documentation on the official OpenCode site. The entry point is OpenCode Console. Related credit and trial notes are collected on the free AI API credits page. We haven't logged in to the Console to try this free tier. The details above come from the official documentation, the pricing table, and the models API.