Fireworks AI has released Ember-1, a specialized model from Fireworks Research built by post-training Moonshot AI's open-weight Kimi K3. Ember-1 learns to produce shorter reasoning traces while preserving task accuracy. This differs from lowering the reasoning effort setting at inference time. According to the Fireworks release post, Ember-1 delivers Kimi K3's quality with about 40% fewer tokens.
Is it deployable? Yes, but only through the Fireworks serverless API as a Research Preview. Fireworks has not released Ember-1's weights, training code, or exact training algorithms, so self-hosting is not an option today.
The Problem: Reasoning Models Think Too Much
Fireworks reports that reasoning models like Kimi K3 sometimes spend more than 90% of generated tokens on internal reasoning. That cost compounds in multi-turn agentic workloads. Each turn replays prior reasoning back to the model, and context grows roughly quadratically with the number of turns. Long traces from early turns get re-read, and re-billed, on every subsequent call.
Fireworks explains that customers wanted K3's coding capability at a lower cost. Turning down K3's reasoning effort did not solve the issue, since lower effort settings sacrificed too much quality. So the team trained the model to reason more efficiently instead.
How Fireworks Research Built Ember-1
Not all of K3's reasoning is waste. Some of it is useful self-reflection, such as revisiting an assumption or reacting to feedback. Ember-1 preserves that behavior while cutting redundant reasoning and unproductive loops.
The training collection spans mathematics, coding, instruction following, conversation, search, tool use, and software engineering. It covers both standalone problems and extended tasks.
via MarkTechPost