MiniMax M3: an open-weight model claiming frontier coding at 1M context

Released June 1, MiniMax M3 is pitched as the first open-weight model to combine frontier coding, a 1M-token context window, and native multimodality in one architecture, with a self-reported 59.0% on SWE-Bench Pro (above GPT-5.5 and Gemini 3.1 Pro, approaching Claude Opus 4.7) at $0.30/M input tokens, weights promised on Hugging Face within ~10 days. Treat the benchmarks with caution — TechTimes flagged them as unverified — but if they hold, it's a serious self-hostable option for repo-scale, long-context agentic coding without per-token frontier pricing.

---

Sources: MiniMax · The Decoder