AI Models & Tiers
Conduit provides access to AI models from Anthropic (Claude) and OpenAI (GPT). Available models depend on your subscription tier.
Model Availability by Tier
Free Tier (2 models)
- Haiku 4.5 (Anthropic) — Fast, efficient for simple tasks
- GPT-5 Mini (OpenAI) — Compact, quick responses
Pro Tier (6 models)
Everything in Free, plus:
- Sonnet 4.6 (Anthropic) — Balanced performance and capability
- Opus 4.5 (Anthropic) — High capability
- Opus 4.6 (Anthropic) — Latest high-capability model
- GPT-5.2 (OpenAI) — Advanced reasoning
Team Tier (9 models)
All models including beta releases.
Token Limits
- Free — 1M monthly, 200K daily, 4K max output
- Pro — 10M monthly, unlimited daily, 8K max output
- Team — 10M/seat monthly, unlimited daily, 16K max output
Model Selection
Switch models per-conversation or set a persistent default in Settings. Your default model is validated against your tier on startup — if you downgrade and your default is no longer available, it falls back to the highest-tier model you have access to.
Your default model preference persists across app restarts and is validated against your current tier.
Backend Proxy
All AI calls route through the Conduit backend for token tracking and tier enforcement. No direct API keys are needed — your Conduit account handles authentication and usage limits.