Fireworks launches Ember-1, matching Kimi K3 quality with 40% fewer tokens
What happened
Fireworks Research 基于 Kimi K3 训了个新模型 Ember-1，把推理时产生的内部思考 token 压掉了 35% 到 50%，但准确率没掉。他们跑了 50 多次训练实验，发现 K3 超过九成的 token 都花在内部推理上，其中很多是多余的。Ember-1 保留了有用的自我纠错，跳过了没产出的思考循环。在多轮 agent 任...
Coverage
Follow the reports to see the story from different sides.
- AI HOT (Curated Pool)PickFireworks launches Ember-1, matching Kimi K3 quality with 40% fewer tokens
Fireworks Research released Ember-1, a model built on Kimi K3 that cuts reasoning tokens by 35–50% while keeping accuracy. Across 7 benchmarks and live A/B tests with two customers, quality held. The team ran 50+ training experiments and found K3 spends over 90% of tokens on internal reasoning, much of it unnecessary. Ember-1 preserves useful self-correction and skips unproductive loops. Savings compound in multi-turn agent tasks where prior reasoning is re-read each turn. The model is live on Fireworks' platform as the first in their own model series.