Fireworks AI Releases Ember-1: A Post-Trained Kimi K3 That Uses About 40% Fewer Tokens
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 keeping task accuracy. This is different 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
Get smart on it
Fireworks AI has released Ember-1, a specialized model created by retraining an open-weight reasoning model to produce shorter internal reasoning traces while maintaining task accuracy. Reasoning models typically spend over 90% of their generated tokens on internal reasoning, which becomes costly in multi-turn interactions where prior reasoning gets replayed and re-billed with each new turn. Ember-1 achieves approximately 40% fewer tokens than its base model according to Fireworks' evaluations and production testing, with the savings coming from more efficient reasoning rather than simply reducing reasoning effort settings. The model is available only through a research preview API and Fireworks has not released its weights or training code, so users cannot run it on their own systems.
Related news
Google Research Moves Federated Learning Into TEEs: Gboard Now Trains With Externally Verifiable Differential Privacy
Google Research has announced a next-generation Federated Learning (FL) system built on Trusted Execution Environments (TEEs). The research team claims externally verifiable central differential privacy (DP) guarantees for FL for the first time. What Problem Does TEE-Based Federated Learning Solve? Google introduced Federated Learning FL in 2017. It powers next-word prediction and Smart Compose on Gboard, reply suggestions in Google Messages, and Smart Text Selection in Android. Earli
Aleph Alpha Releases Kolibri: A 78.1B Open-Weight English-German MoE Model With Only 3.46B Active Parameters
Aleph Alpha has released Kolibri, an open-weight Mixture-of-Experts (MoE) language model built for German and English. Kolibri has 78.1B total parameters but activates only 3.46B, or 4.4%, per token. It accepts up to 1,048,576 tokens of context, lets users set reasoning effort per request, and ships under the Apache 2.0 license on Hugging Face. The target is sovereign deployment in regulated sectors such as public administration, industry and aerospace. Is it deployable? Yes. The FP8 checkpo
