Taskade Genesis routes every prompt, agent run, and automation step through 15+ frontier models from OpenAI, Anthropic, Google, and leading open-source providers. The picker shows the credit cost next to each model so you always know what you are spending before you commit. Published apps run for free. Visitors never burn your credits.
TL;DR: Credits cover building. Published Taskade Genesis apps are free to run, no matter how many visitors. Lightweight models cost a few credits per use, deeper reasoning models cost more, and the picker shows the live price before you run a step. Auto mode picks for you. Free plan gets 3,000 one-time credits; paid plans (Starter $6, Pro $16 for 10 users, Business $40, all annual) include monthly allowances. Compare plans →
Model Tiers
The picker groups models by what they are great at. Cost rises with the depth of reasoning, not the brand name.
| Tier | Great picks | Credit cost | Best for |
|---|---|---|---|
| Lightweight & fast | Google Gemini Flash, Anthropic Claude Haiku | Cheapest | Quick iterations, simple text, high-volume agent calls |
| Standard | OpenAI GPT family, Anthropic Claude Sonnet, Google Gemini Pro | Mid | General reasoning, code, balanced quality |
| Open-source value | Qwen, Kimi, DeepSeek, MiniMax, GLM | Low to mid | Cheap routine steps, bulk routing, cost-sensitive automations |
| Long-context | Anthropic Claude Opus, Google Gemini Pro (extended) | Higher | Long documents, deep analysis |
| Deep reasoning | OpenAI reasoning models | Highest | Math, multi-step planning, complex logic |
Taskade adds new models as they ship. Use Auto mode when you don't want to choose. Auto picks the right tier for the job so you only pay for depth when you need it.
Credits by Plan (annual billing)
| Plan | Price | Best for |
|---|---|---|
| Free | $0 | Try Taskade Genesis (3,000 one-time credits) |
| Starter | $6/mo | Solo builders |
| Pro | $16/mo | Teams up to 10 users |
| Business | $40/mo | Growing companies |
| Max | $200/mo | Power users with high AI usage |
| Enterprise | $400/mo | Larger teams, SSO, security review |
Tips for Optimizing Credits
- Use lightweight models for cheap iterations and high-volume agent calls
- Reserve reasoning models for math, planning, and hard logic
- Publish your app the moment it's stable — published apps cost zero credits to run
- Let Auto mode route automatically if you're unsure
See also
- Custom AI Agents — Per-agent model selection
- Full pricing details
The Key Insight: Published Apps Are Free to Run
Build once, run forever. Credits are consumed while you're iterating on an app in the builder — not when visitors use the published version. A Starter plan ($6/mo annual) gives you credits to build and iterate on your apps, and those apps can serve unlimited visitors afterwards.
Related guides
- AI Models. Which of the 15+ frontier models to pick.
- Publishing. Publish to lock in the free-runtime behavior.
- Taskade Pricing. Plan comparison.
- Buy Extra Credits. Pay-as-you-go top-ups.