Ember-1 Cuts Kimi K3 Token Costs

Ember-1 targets the real bottleneck in agentic AI: reasoning cost, not just model quality.

• –Shorter reasoning traces could significantly reduce expenses in long-running coding and tool-use agents.
• –Fireworks reports comparable quality across seven benchmarks, customer A/B tests, and internal coding workloads.
• –The model supports a 1M-token context window, function calling, and image input, broadening its production use cases.
• –Research-preview status and company-run evaluations mean developers should validate quality and reliability on their own workloads.
• –Serverless pricing is listed at $3 per million input tokens and $15 per million output tokens.