Gravity puts ads under three LLMs and cuts token prices by up to 25%
Gravity's Inference.net launch lists Kimi at 20% off and DeepSeek and GLM at 5% off, with ads appended to about 4.2% of answers.
By Ryan Merket · Published
Primary source: x.com
Why it matters
This is the first attempt to wire an advertising subsidy directly into per-token inference pricing rather than a consumer chat interface. If advertisers fund even a 5% to 20% discount at scale, inference gateways gain a pricing lever that does not depend on cheaper hardware or smaller models, and AI wrapper startups must decide whether ad-appended answers are acceptable inventory. It also puts Gravity, with $38.5 million raised, into direct competition with OpenAI's advertising business for the same advertiser budgets.
Gravity launched ad-supported versions of three open models on the Inference.net gateway on November 9th, selling access at prices it says run up to 25% below standard per-token rates. The lineup is Moonshot AI's Kimi K3.1 Pro, DeepSeek's V4.1 Flash and Z.ai's GLM 5.3, according to Gravity's launch thread on X. Gravity says ads are appended to the bottom of an answer about 4.2% of the time, only when relevant to the response, and that the ads do not change the model output itself.…
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.