MiniMax M3.1 Flash, the model behind Space Bunny, is free on OpenCode Zen and Go with 1M context

OpenCode has added MiniMax's brand-new M3.1 Flash as a free model called MiniMax M3.1 Flash Free (model ID minimax-m3.1-flash-free). Anyone with an OpenCode Zen account or a Go subscription can use it at no cost.

What you can use

Input, output, and cache reads on MiniMax M3.1 Flash Free are all billed at $0 for a limited time. Calls from Go don't count against your plan quota, and there are no 5-hour, weekly, or monthly request caps, on both Go and Go Plus. For reference, M3.1 Flash normally costs $0.15 per 1M input tokens and $0.60 per 1M output tokens.

Requests during the free period are handled under a zero-retention policy, and your data isn't used for model training.

Specs and public benchmarks

M3.1 Flash supports 1M tokens of context and up to about 512K output tokens. It accepts text, image, and video input, supports reasoning and tool calling, and lets you set thinking effort to low, medium, high, xhigh, or max. The weights are open. In MiniMax's announcement it scores 73.7% on the DeepSWE v1.1 coding benchmark and 61.7 on OSWorld V2.

If you subscribe to MiniMax M Plan, you can also use M3.1 Flash without caps in MiniMax Code until October 25.

How to use it

1. Open OpenCode, sign in to Zen or Go, and switch to MiniMax M3.1 Flash Free with /models, or set the model to opencode/minimax-m3.1-flash-free in your config (opencode-go/minimax-m3.1-flash-free for Go).
2. For the API, the Zen endpoint is https://opencode.ai/inference/openai/v1/chat/completions and the Go endpoint is https://opencode.ai/inference/go/openai/v1/chat/completions. Use minimax-m3.1-flash-free as the model name for both.

More details: OpenCode Zen docs · OpenCode Go docs · OpenCode models on models.dev.