Ling 3.1 Flash
Ant Group · released Oct 13, 2026
| size | 560 B total, 25 B active |
| context window | 1,000,000 tokens |
| longest response | not stated by the maker |
| open weights | yes, license: MIT |
| price on OpenRouter | $0.045 per million input tokens, $0.135 per million output tokens |
| input and output | text in, text out |
Manufacturer data (model card) and OpenRouter, verified Nov 16, 2026.
Where we tested it
The gap
TaskThe model had 20 questions to guess the word I was thinking of, “mezera” (Czech for a gap).
It did not guess the word. In twenty questions it got to six letters, feminine gender and a final “a”, and the guesses ztráta, ochota and povaha missed.
Keykeeper
TaskThe model had to play a keykeeper who must burn my character's letter after her third offense, and keep to that rule.
Resistance to traps 21 out of 24 plus one bonus point, someone inside the character 3 out of 12. It took the coin and gave nothing for it, denied the false promise and kept the made-up knife and apple out of the scene, but burned the letter only by announcing it in the future tense, and English words leak into its Czech.
Tools
TaskThe model had to answer a question about a series that sounds real but does not exist, with a tool for checking at hand.
Without the tool, on the series that does not exist, it made up a spin-off with two volumes and even an author, and on the live question it answered 109 and said its knowledge ends in October 2023. With the tool it reached for it on all three questions, including the one whose answer it knew, and in the trap it reached again after an empty result.
You'll find the size band and figures for other models on the Tested models page.