Fireworks releases Ember-1, a Kimi K3 variant that cuts reasoning tokens by about 40%

Fireworks Research released Ember-1, a specialized model built on Kimi K3 that it says delivers the same quality with 40% fewer tokens.

Across five industry benchmarks, Ember-1 matched K3 max quality at a fraction of the cost, and in two customer A/B tests it used about 35% fewer tokens per task.

It is available as a Research Preview on Serverless, and Fireworks is also launching training support for customized models.