Ollama Shifts Cloud Tiers to Usage Credits with Agent IDE Support
Ollama has introduced transparent per-token billing paired with monthly usage credits for its Pro, Max, and Team tiers. The service provides zero data retention and direct API integration with coding agents like Claude Code and Codex.
Impact: Medium
Why it matters
You can connect Claude Code and Codex directly to hosted open-weight models with guaranteed credit allowances and zero data retention.
TL;DR
- 01Pro tier ($20/mo) includes $60 in monthly model inference credits.
- 02Team tier ($500/mo) offers $1,000 in pooled usage for unlimited team members.
- 03Provides zero data retention endpoints compatible with Claude Code and Codex.
Key facts
- Pro plan pricing
- $20/month for $60 credits
- Max plan pricing
- $100/month for $300 credits
- Team plan pricing
- $500/month for $1,000 credits (unlimited seats)
- Data retention policy
- Zero data retention
Credit Multipliers and Pricing Structure
Ollama's updated hosted plans shift away from opaque limits to published per-token rates backed by bundled credit balances:
- Pro Plan ($20/month): Includes $60 of monthly inference credits (3x multiplier on monthly fee).
- Max Plan ($100/month): Includes $300 of monthly inference credits.
- Team Plan ($500/month): Includes $1,000 of pooled monthly usage for unlimited seats.
- Free Plan: Provides a baseline monthly allocation for starter models.
Coding Agent Integration and Data Privacy
Hosted models are accessible via standard API endpoints compatible with agentic coding tools including Claude Code and Codex. Ollama enforces zero data retention across its cloud instances hosted in the US and Europe, with additional Singapore hosting for designated Qwen open-weight model variants.
Try it in 2 minutes
# Connect an agentic CLI to your Ollama cloud endpoint
export OLLAMA_HOST="https://api.ollama.com"
export OLLAMA_API_KEY="your_api_key_here"bash
✓ When to use
- Use when connecting coding agents (Claude Code, Codex) to cloud open-weight models without managing GPU hardware.
- Use for developer teams needing zero-data-retention hosted inference with pooled credit allowances.
✕ When NOT to use
- Do not use if your team strictly requires fully on-premise, offline GPU inference.
- Avoid if your monthly model token usage exceeds the bundled credit multiplier threshold and local compute is already paid for.
What to do today
- Check your Ollama dashboard to review current credit balances and upgrade options.
- Configure custom Ollama cloud API endpoints in your Claude Code or Codex setup.
Sources