Skip to main content

Using Fireworks AI With Roo Code

Fireworks AI is a high-performance platform for running AI models, offering fast, cost-effective access to a wide range of state-of-the-art open-source language models. Built for speed and reliability, Fireworks AI provides serverless deployment options with OpenAI-compatible APIs and context windows up to 256,000 tokens.

Website: https://fireworks.ai/

Getting an API Key​

  1. Sign Up/Sign In: Go to Fireworks AI and create an account or sign in.
  2. Navigate to API Keys: After logging in, go to the API Keys page in the account settings.
  3. Create a Key: Click "Create API key" and give your key a descriptive name (e.g., "Roo Code").
  4. Copy the Key: Copy the API key immediately and store it securely. You will not be able to see it again.

Supported Models​

Fireworks AI offers several high-performance models:

Kimi K2​

  • accounts/fireworks/models/kimi-k2-instruct (Default)
    • 1 trillion total parameters with 32B activated params
    • 128K context window
    • Optimized for agentic capabilities
    • $0.60/M input, $2.50/M output

Qwen3 Series​

  • accounts/fireworks/models/qwen3-235b-a22b-instruct-2507

    • 256K context window
    • Competitive with closed-source models
    • $0.22/M input, $0.88/M output

  • accounts/fireworks/models/qwen3-coder-480b-a35b-instruct

    • 256K context window
    • Specialized for coding tasks
    • $0.45/M input, $1.80/M output

DeepSeek Series​

  • accounts/fireworks/models/deepseek-r1-0528

    • 160K context window
    • Advanced reasoning with reduced hallucination
    • Function calling support
    • $3.00/M input, $8.00/M output

  • accounts/fireworks/models/deepseek-v3

    • 128K context window
    • 671B total parameters with 37B activated
    • $0.90/M input, $0.90/M output

Configuration in Roo Code​

  1. Open Roo Code Settings: Click the gear icon () in the Roo Code panel.
  2. Select Provider: Choose "Fireworks AI" from the "API Provider" dropdown.
  3. Enter API Key: Paste your Fireworks AI API key into the "Fireworks AI API Key" field.
  4. Model Selection: The default model (Kimi K2) is automatically selected. You can change it from the model dropdown if needed.

Model Selection Guide​

Choose models based on your needs:

ModelBest ForContextPrice
Kimi K2General tasks, balanced performance128KMid-range
Qwen3 235BCost-effective general use256KBudget-friendly
Qwen3 CoderCode generation and debugging256KMid-range
DeepSeek R1Complex reasoning, function calling160KPremium
DeepSeek V3Strong general performance128KBalanced

Tips and Notes​

  • Cost-Effective: Fireworks AI offers significantly lower pricing than proprietary models while maintaining competitive performance.
  • Large Context Windows: Most models support 128K-256K tokens, suitable for processing large documents and maintaining extended conversations.
  • OpenAI Compatibility: The provider uses an OpenAI-compatible API format with streaming support and usage tracking.
  • Rate Limits: Fireworks AI has usage-based rate limits. Monitor your usage in the dashboard and consider upgrading your plan if needed.
  • Text-Only: All models are text-only without image support or prompt caching capabilities.
  • Default Temperature: Uses 0.5 temperature by default for balanced creativity and consistency.
  • API Keys: Stored locally on your machine for security.
  • Pricing: See the Fireworks AI pricing page for current rates. Prices shown are per million tokens.