Novita AI (Ling-3.0-Flash-Fin, free for a limited time)

Ant Group's Ling-3.0-Flash-Fin (a finance-enhanced MoE model from the Ling/InclusionAI family) is free on Novita AI for a limited time — 30 days — with both input and output priced at $0/M tokens

The full model name is inclusionai/ling-3.0-flash-fin, part of Ant Group's Ling (InclusionAI/Ling) family — a finance-focused enhancement built on Ling-3.0-Flash, with 124B total parameters and roughly 5.1B active parameters, specializing in multi-step financial reasoning and long-chain execution for tasks like financial-report analysis, investment research, and valuation. It has a 256K (262,144) context window, 32K max output, and supports tool/function calling. Launched 2026-08-27; Novita's official model page currently confirms pricing of $0/M tokens for both input and output, available through both Novita's own Serverless API and OpenRouter (provider: NovitaAI), with the OpenRouter side marked as rate-limited.

⚠️The official basis for the free period is Novita's official Twitter/X post stating 'free for 30 days' (the model launched 2026-08-27), but no explicit calendar end date has been officially published — this entry deliberately doesn't fill in an inferred date (better to leave it blank than guess). For reference, the non-finance version in the same family, Ling-3.0-Flash, has previously had its 'limited-time' promotion extended before; whether this one will be extended is unknown — check the model page to confirm pricing is still $0 before use. No credit card required to register; direct access from mainland China requires a proxy/VPN. Novita's official documentation separately confirms that new users get a voucher trial quota on signup (amount unspecified; third-party channels report roughly $0.5, not independently verified against the official page — for reference only, not the main subject of this entry).