Models & Providers

Blaze supports multiple AI providers and their full suite of models. In the Models section, you can configure which "brain" powers your sessions and monitor performance metrics.

Provider Management

Switch between supported providers and see their status at a glance:

• Anthropic: Claude 3.5 Sonnet, Opus, and Haiku.
• Google: Gemini 1.5 Pro and Flash.
• OpenAI: GPT-4o, o1-preview, and o3-mini.

Model Tiers

Blaze categorizes models into three tiers to help you balance capability and cost:

• Flagship (Purple): Most capable models (Opus, o1) for complex architecture and reasoning.
• Balanced (Blue): The sweet spot for speed and intelligence (Sonnet, GPT-4o).
• Fast (Green): Lightning-fast response times for simple tasks and formatting (Haiku, Flash).

Context & Cost

• Context Windows: See the max token limit for each model (e.g., 200k for Claude).
• Cost Tracking: Monitor your estimated spend per million tokens based on current provider rates.

Next: Engines & CLI Adapters.