MiniMax M3 hits Cloudflare AI Gateway

// 1h agoMODEL RELEASE

MiniMax M3 hits Cloudflare AI Gateway

Cloudflare has made MiniMax's M3 open-weight model available on AI Gateway, offering native multimodal capabilities and a 1-million-token context window. To celebrate the integration, all gateway requests with context windows up to 512K tokens receive a 50% discount during the launch week.

// ANALYSIS

Cloudflare's rapid integration of MiniMax M3 signals a highly competitive landscape for open coding and multimodal models where distribution networks like AI Gateway can make or break model adoption.

–The 1M token context window combined with MSA (MiniMax Sparse Attention) positions M3 as a serious open alternative to closed models like GPT-4o or Gemini 1.5 Pro.
–Offering a 50% discount for shorter contexts (≤ 512K) during the launch week lowers the barrier for developer testing and helps seed the developer ecosystem early.
–Cloudflare AI Gateway continues to establish itself as an essential developer hub by serving as the bridge to easily integrate and manage complex multimodal models.

// TAGS

minimaxm3cloudflareai-gatewayllmmultimodalopen-source

DISCOVERED

1h ago

2026-06-01

PUBLISHED

1h ago

2026-06-01

RELEVANCE

8/ 10

AUTHOR

CloudflareDev

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE27m ago

Ultracode demands stacked Claude Max subscriptions

A developer has evaluated the practical cost of running Claude Code's high-effort "Ultracode" mode, stating that the feature requires at least two Claude Max subscriptions to be viable for professional workflows. Although they spend $600 monthly on three Claude Max subscriptions to sustain their agentic coding tasks, they consider it the best money they spend due to the substantial productivity benefits.

LAUNCH34m ago

Google launches new agent developer tools for the Gemini Enterprise Agent Platform at I/O 2026, featuring Antigravity 2.0, Managed Agents API, and Agent Studio to streamline secure, production-ready AI agent deployments.

At Google I/O 2026, Google introduced a powerful suite of developer tools for the Gemini Enterprise Agent Platform to simplify the process of building, testing, and deploying autonomous enterprise AI agents. The new suite accommodates both low-code and code-first workflows: Agent Studio provides a visual workspace for non-technical teams, while the Agent Development Kit (ADK) 2.0 and Antigravity 2.0 (formerly the Gemini CLI) offer code-first developer environments for complex, multi-agent systems. Crucially, the new Managed Agents API allows developers to launch fully sandboxed, secure, Google-hosted agents with a single API call, resolving key infrastructure and security issues for production deployments. Integrated with Gemini 3.5 Flash and the open Model Context Protocol (MCP), these tools bridge the gap between local prototyping and secure cloud deployment.

UPDATE36m ago

Lovable projects ship with TanStack Start

AI full-stack application builder Lovable now builds and ships all new projects using TanStack Start as the default meta-framework. This integration provides developers with full-stack React capabilities out-of-the-box, enabling flexible per-route routing options (SSR, SSG, or CSR) alongside built-in support for server functions.

MiniMax M3 hits Cloudflare AI Gateway