InclusionAI's Ling 3.0 Flash Fin is a free 124B MoE on OpenRouter
A 262K-context MoE you can wire into Cline or Kilo Code for $0 — plus Vercel's new Slack-bot template and a CLI that manages your domains.
Copy markdown
A free 124B MoE, OpenAI-compatible
InclusionAI's Ling 3.0 Flash Fin is live free on OpenRouter: 124B params, 5.1B active per token, and a 262K context window, hosted by NovitaAI. It speaks the OpenAI API, so you can point Cline, Kilo Code, or any custom-base-URL client at it and pay nothing.
The catch: free API isn't open weights
Free access isn't the same as open weights — there's still no Hugging Face model card, and a single host (NovitaAI) means the free tier carries rate limits and could disappear. It's also tuned for long-horizon finance and agent workflows, though InclusionAI says it keeps general reasoning, coding, and math.
Elsewhere: a Slack bot on Claude Managed Agents
Vercel shipped a template pairing Claude Managed Agents with its Chat SDK — a Slack research bot that keeps a persistent session per thread. It's a clean weekend build if you want a team bot with memory instead of stateless one-shots.
Elsewhere: Vercel's CLI now manages your domains
The Vercel CLI can now inspect and update DNS records, renew domains, and configure projects and team members — the kind of dashboard clicking you can finally script into CI.