[KongchangAI]
ModelNimotron

Nimotron 3 Ultra

A low-cost large language model that ranks near the bottom of the code-to-video generation benchmark and is positioned around extremely low inference cost.

Timeline (last 90 days)

Oct 27

Nimotron 3 Ultra came last in the Code2Video Benchmark, 1232 points behind the first place, with an average task cost as low as 3 cents.

Unverified50%
Oct 27

The average task cost of Claude Opus 5.5 and Nimotron 3 Ultra differs by about 100 times ($3.08 vs 3 cents).

Unverified50%

All Facts (2)

Source Articles