Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-08-26 · high
The free routing through GMI Cloud closes September 6, and the model IDs stop returning entirely after that.
Why now: Vercel's changelog dates the offer through September 6, leaving eleven days from August 26 to test before the free tier disappears.
Vercel opened free access to MiniMax M3 and M2.7 on its AI Gateway, routed through GMI Cloud, through September 6. The model IDs, minimax/minimax-m3-free and minimax/minimax-m2.7-free, work like any other AI Gateway route and cost nothing until the window closes.
This is useful for anyone testing MiniMax against a paid default model. Swap in the free ID, run the same prompts, and compare MiniMax against whatever you're already paying for, no GMI Cloud account and no card on file, per Vercel's changelog.
After September 6 the free IDs stop returning results. Vercel's post doesn't say whether a paid MiniMax route through AI Gateway continues after that date or what GMI Cloud charges once the free tier ends.
This free window looks like a GMI Cloud funnel more than a pure benchmarking gift. The real test is whether anyone keeps paying to route through GMI Cloud once the free IDs disappear on September 6, not whether MiniMax M3 benchmarks well.
Vercel's changelog dates the offer through September 6, leaving eleven days from August 26 to run a comparison before the free tier ends.
Each link below shares sources, entities, or timing with this story.
Satya Nadella said companies routing everything through a single proprietary lab may not survive. His argument: you hand that lab your most sensitive business context, and the lab can turn it against you as a competitor. His prescription is an orchestration layer — keep the ha...
OpenRouter and Vercel's AI Gateway both list $2.50/M input and $15.00/M output, down from $5.00/$30.00, while OpenAI's own API docs show standard pricing. Vercel's changelog dates it precisely: Aug 17 through Sep 18, all requests through AI Gateway. Gateway-only and time-boxed...
vercel ai-gateway coding-agents setup routes Claude Code, Codex, OpenCode, Pi, Cline, Cursor, Hermes, Kilo Code, and OpenClaw through AI Gateway, consolidating spend, traces, tokens, and model attribution into one dashboard with per-key budgets (--budget 500 --refresh-period m...
Announced August 7: Hermes can use AI Gateway as its inference layer for 200+ models with no token markup and per-request dashboard visibility, and execute shell commands inside an isolated Vercel Sandbox microVM instead of on your machine, with Node.js 24/22 and Python 3.13 a...
GPT-5.6 Luna went to $0.20 input / $1.20 output per million tokens on July 30. That's an 80% cut. Terra dropped 20%. Luna's input now undercuts Gemini 3.1 Flash-Lite ($0.25/$1.50) and sits at one-fifth of Claude Haiku 4.5's $1 input. Simon Willison covered the announcement and...
I've spent the last year assuming that if I wanted real agentic coding quality, I paid for a closed model. That assumption took a hit on June 1. MiniMax shipped M3 with a new sparse-attention architecture (they call it MSA) that handles up to 1M tokens at roughly 9x prefill an...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.