Use MiniMax Models in Big-AGI.

Running MiniMax in Big-AGI

Connect MiniMax's OpenAI-compatible API with your own key, or route the M-series through OpenRouter. Either way, Big-AGI adds no markup and no intermediary: billing runs directly between you and MiniMax or OpenRouter, and your keys stay in your browser.

• Your key, your billing. Usage is billed by MiniMax or OpenRouter to your account. Big-AGI does not meter or charge for model usage.
• Frontier value, tuned for agents. The M-series pairs a million-token context window with aggressive pricing, much of it open weights, and from M2 on the models are built for coding and multi-step tool use.
• A catalog that just works. MiniMax has no models-listing API at all. Big-AGI recognizes the MiniMax hostname and serves an always-current, hand-maintained catalog with real pricing instead of hammering a dead endpoint. Setup stays one paste of a key.

Why run MiniMax here?

MiniMax's most visible consumer product is Talkie, a character-chat app, not a workbench for the M-series models. Beam is that workbench: it runs MiniMax next to GPT, Claude, and Gemini on the same prompt, so you can see whether its long-context read or its agentic plan holds up against the rest of the field before you rely on it. You also get parameters no character app exposes (temperature, system prompt, per-turn model swaps), and a key that stays in your browser.

Your keys and your data

Turn on Direct Connection and the browser calls MiniMax, or OpenRouter, directly, bypassing the Big-AGI server, whenever your key is client-side and the provider allows it. Your keys stay in your browser. Chats are stored locally first and sync only if you turn it on. The AI Inspector opens on any message to show the exact request, the token counts, and a cost estimate for that call.

MiniMax in Beam

Run MiniMax in parallel with Claude, GPT, and Gemini on the same prompt, then reach for Fusions: several strategies that combine, cross-check, and synthesize the parallel answers, which beats just picking the single best one. Parallel runs use more tokens than a single chat.