Novita AI

On August 14, 2026, Novita AI published a Model Deprecation Notice announcing the retirement of kwaipilot/kat-coder-pro from the Novita platform on August 31, 2026 at 15:59 UTC. No replacement model was listed for this deprecation. The notice instructs developers whose applications use the affected model to stop sendin This is a single-model deprecation with no alternative offered, reducing its direct migration utility, but it provides a verifiable record of the model lifecycle change. The page was last modified on August 27, 2026.

Novita AI published a free LLM API comparison on August 18, 2026, detailing concrete free-tier mechanisms across six providers. The company's own free model is a referral credit of $10 to each of the referrer and the new account per successful signup, capped at $500 total, with no per-minute rate limit on the full serv The article emphasizes that no provider on the list offers unlimited free LLM API access; every free tier is bounded by request rate, daily cap, or dollar credit. Credit-based free tiers (Novita AI, Hugging Face) avoid per-minute rate caps, while rate-limited tiers (OpenRouter, Groq) impose model-specific limits. The p

Novita AI published a first-party model deprecation notice on August 12, 2026 announcing that the inclusionai/ling-3.0-tiny model will be retired from the Novita platform on August 13, 2026 at 15:00 UTC. The notice identifies it as a time-limited free trial model released on August 5, 2026, and instructs developers to The short notice window (one day between announcement and retirement) and the lack of a designated replacement make this a concrete operational update that developers integrating with the Novita API need to act on. Because the page is the official Novita documentation changelog, the retirement timestamp and model ident

Novita AI has partnered with Hugging Face to provide instant AI model deployment for over five million developers on the platform, according to an April 2026 announcement. The collaboration introduces a "Deploy on Novita" feature that lets developers turn models into production-ready APIs without managing infrastructur Novita AI claims time-to-first-token as low as 50 milliseconds and cost savings of up to 50% compared to most inference endpoints, with the platform supporting over 120 large language and multimodal models through a single API. COO Junyu Huang framed the integration as eliminating complex deployment steps so developers