Sparse Mixture-of-Experts (MoE) autoregressive transformer with test-time reasoning, post-trained from Kimi K3
MoEPro
Ember-1 is Fireworks Research's first custom post-trained model: a Kimi K3 derivative re-trained to emit roughly 40% fewer reasoning tokens at comparable quality, shipping with a 1,048,576-token context window, text+image input, and $3/$15 per-million-token pricing. It is a research preview, not a new architecture — and its only published third-party scores are the Vals AI real-world agentic suites (Vibe Code Bench v1.1 83.9%, CyberBench v1.1 69.5%, Finance Agent v2 51.9%, Terminal-Bench 4.0 ~20.7%).
Vibe Code Bench v1.1:83.9%
1.05M ctx
Specialized Modality
