Novita AI delists Ling 3.0 Tiny, emptying its free-LLM shelf after 3 days
Novita AI delisted Ling 3.0 Tiny, the free LLM added 3 days earlier, leaving no $0 language model on its rate card as the catalog eased from 177 to 174 listings.
Ling 3.0 Tiny listed at $0/M input and output under a Time Limited Free tag, added 2026-08-11 as the free-launch slot; 177 models on the serverless catalog.
Ling 3.0 Tiny no longer appears on the pricing page or model catalog; no free LLM listed on the rate card; catalog count 174 models.
Proof of change
Novita AI's pricing pages, as we captured them on two dates.
Showing the whole page as captured — scroll either panel, or open it at full size.
- 2 pages had no counterpart in the earlier capture and are not shown.
Both images are our own captures, taken on the dates shown. The pixels are unmodified — nothing is retouched, and where a region is highlighted the marker is drawn over the image, not into it. Open either capture to see it at full resolution.
Novita AI’s serverless catalog lost its only free-tier language model for the second time in a month. Ling 3.0 Tiny — added on 2026-08-11 at $0/M input and output under the same “Time Limited Free” tag Novita uses to launch new small models — is gone entirely from both novita.ai/en/pricing and novita.ai/en/models as of the 2026-08-14 capture. Unlike Tencent’s Hy3, which graduated from free to a paid per-token rate on 2026-07-21, Ling 3.0 Tiny was delisted outright rather than repriced, the shortest lifespan of any free listing tracked on this rate card so far.
The overall model count eased from 177 to 174 in the same capture, with no other confirmed per-token rate changes: the previously tracked LLM, GPU-instance, dedicated-endpoint, bare-metal, and Agent Sandbox prices were all unchanged. A “Limited Time 50% Off” tag previously shown next to Qwen3.7-Max also no longer appears on the rate table, though its $1.25/M input and $3.75/M output rate held steady rather than reverting higher.
Ling 3.0 Tiny — the free LLM added 2026-08-11 — was removed outright from the pricing page and model catalog rather than converted to a paid rate, the shortest-lived free listing tracked on this rate card. The catalog eased from 177 to 174 models; Qwen3.7-Max's 'Limited Time 50% Off' tag also disappeared from the rate table though its $1.25/$3.75 rate held steady. All other tracked per-token, GPU, dedicated-endpoint, bare-metal, and Agent Sandbox rates were unchanged.