Ember-1
Ember-1 is Fireworks AI’s specialized Kimi K3-derived reasoning model, offered through its serverless API and playground. The 2.78T-parameter mixture-of-experts model supports image inputs, function calling, and a context advertised as 1040K tokens. It is optimized for shorter reasoning traces: the publisher reports roughly 40% fewer generated tokens with comparable results on its own evaluations. This is a provider evaluation claim rather than a guarantee of latency or quality in every agent workload. Ember-1’s September 23 launch establishes public availability, while the catalog’s September 22 creation timestamp is earlier metadata. Public model weights are not offered.
2026-09-23
2.78T total
Mixture-of-Experts Transformer
Proprietary
Specifications
- Parameters
- 2.78T total
- Architecture
- Mixture-of-Experts Transformer
- License
- Proprietary
- Context Window
- 1,040,000 tokens
- Type
- multimodal
- Modalities
- textimage
Benchmark Scores
Advanced Specifications
- Model Family
- Ember
- Finetuned From
- Kimi K3
- API Access
- Available
- Chat Interface
- Available
- Variants
- accounts/fireworks/models/ember-1
Capabilities & Limitations
- Capabilities
- reasoningcodingimage inputfunction callinglong context
- Known Limitations
- Provider token-efficiency results depend on its evaluation setupNo public weights or supported fine-tuning endpointCatalog rounds context to 1040K; an exact binary limit is not asserted
- Notable Use Cases
- coding agentstool-based workflowsdocument reasoning
- Function Calling Support
- Yes
- Tool Use Support
- Yes
