Fireworks AI

Fireworks AI runs open models from many vendors behind one API. In LM15 its provider is fireworks.

Provider	fireworks (litellm spelling: fireworks_ai/)
Key	FIREWORKS_API_KEY, from app.fireworks.ai/settings/users/api-keys
Model names	Fireworks’ own ids: fireworks:accounts/fireworks/models/deepseek-v4p1-flash
Needs	Python 1.1.0, TypeScript 1.0.0-rc.2, Rust 1.0.0-rc.2, Go v1.1.0-rc.2 or later

Chat, streaming, tools (including a forced tool choice), structured output and the model list work as with any provider. Measured against Fireworks on September 26, 2026:

• Reasoning off is refused by Fireworks, with a clear error, on models that can’t stop reasoning (gpt-oss, GLM-5.3). On models that can, such as DeepSeek V4.1, it works.
• Reasoning stays out of the answer. LM15 sends a model’s earlier reasoning back in the field Fireworks reads, so a tool conversation keeps its train of thought without the reasoning showing in the text.
• Images inside a tool result reach the model.
• Some Fireworks models reason by default, even for a one-word answer; set a low effort for cheap calls.