Intel / post

Ling-3.0-flash-VL Ends Free Tier on OpenRouter, Now Paid

Source: x.com
Date: Published Sep 23, 2026

Context

Ling-3.0-flash-VL, a model from InclusionAI, stopped being free on OpenRouter on September 23 at 9:00 AM Pacific time. Ant Ling's announcement lists limited-time discounted prices per million : $0.021 for input, $0.063 for output and $0.0042 for cache hits.

OpenRouter sells access to many providers' models through a single API, so the change affects developers who call Ling-3.0-flash-VL there rather than running it on their own hardware. Output tokens cost three times as much as input tokens. The cache-hit price applies when input the provider has already processed, such as a repeated prompt prefix, is reused, and it is one fifth of the normal input price, which favours applications that send the same long instructions with every request.

Running the model yourself remains an option: InclusionAI has published quantized fp8 weights for the model on Hugging Face. The post's text calls the OpenRouter prices a limited-time discount but gives no end date and does not say what the price will be afterwards.

Terms in this piece · Glossary

token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
quantization — Shrinking a model by storing its numbers less precisely — like rounding — so it runs faster and fits on smaller hardware, at a small quality cost.

Read the source x.com