Minimax M3 is now available on Telnyx Inference

MiniMax M3 is the first open-weight model to combine frontier coding and agent capabilities, a 1M-token context window, and native multimodal understanding. It's now available on Telnyx Inference, hosted on our own B300 GPU infrastructure.

Powered by MiniMax's Sparse Attention (MSA) architecture, M3 supports up to 1M tokens with a guaranteed minimum of 512K. M3 scores 83.5 on BrowseComp, surpassing Opus 4.7 (79.3).

M3's massive context window processes entire codebases, long documents, and complex workflows in a single conversation. In testing, M3 autonomously optimized a CUDA kernel over 147 benchmark submissions with zero human intervention, achieving a 9.4x speedup. It also independently reproduced an ICLR 2025 Outstanding Paper over 12 hours, producing 18 commits and 23 experimental figures.

What's new:

• 1M-token context window (guaranteed minimum 512K) powered by MSA architecture
• Native multimodal understanding (text, image, and video)
• Hosted on Telnyx-owned B300 GPU infrastructure

We're proud to offer the latest open-weight models as they're released. MiniMax serves 250M+ users globally and 200,000+ enterprise customers. Available through our OpenAI-compatible API.

Get started with MiniMax-M3 on Telnyx.