Ember-1: Fireworks Research's Kimi K3 derivative with 40% fewer reasoning tokens
Fireworks Research's first in-house model: a Kimi K3 post-train that matches K3 quality with roughly 40% fewer reasoning tokens.
Why it matters
First specialized model from Fireworks Research, showing post-training can cut reasoning cost without quality loss; research preview that becomes permanent on community demand.
What you could build with it
A coding-tools team could offer Ember-1 as a budget tier for high-volume agent loops: benchmark cost-per-completed-task vs K3 on their own traffic, then pass the token savings through as a cheaper plan tier.
Does it hold up?
Too early to judge: Fireworks reports 92.2% SWE-bench Verified, 82.0 Terminal-Bench 2.1 and two customer A/B tests, but all figures are vendor-run and no independent evaluation exists yet.
Built with Ember-1
- Ember-1 from Fireworks now available on AI Gatewayarticle · Vercel added fireworks/ember-1 to AI Gateway for use in coding agents with a 1M-token context window.
- HN discussion: Ember-1hackernews · 365-point front-page thread discussing the token-efficiency claims and research-preview terms.
Learn more
First spotted on article: source.
More AI models
ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71DeepSeek V4-Flash official API: public beta with upgraded agent capabilities, Responses API and Codex supportDeepSeek's official V4-Flash is now a public beta API with massively upgraded agent capabilities - benchmark scores…models · JEV 0.69Mistral Forge: enterprise platform for training and continuously improving proprietary modelsMistral AI launched Forge (Sep 28) — a full-lifecycle model training platform (pre-training, SFT, DPO/ODPO, RL…models · JEV 0.69kev: open, trainable Jev-like family of small decision models on Qwen3.5/3.8A family of small decision models (0.8B to 27B) built on Qwen3.5/3.8 that reproduces Jev's prefill-only typed-decision…models · JEV 0.68GPT-Synopsys: OpenAI and Synopsys multi-year deal to build an AI model that operates EDA toolsOpenAI and Synopsys signed a multi-year agreement on Sep 30 to jointly develop GPT-Synopsys, a specialized model that…models · JEV 0.66
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.