Many agentic IDEs and CLIs provide free or paid access to pro-grade models such as Claude Sonnet/Opus, Gemini 2.5 Pro, and GPT o3.
Vendors typically describe limits using tokens, credits, or requests, but that doesn’t always match the actual coding hours before hitting limits. I'd love to hear your real-world experiences.
Reply using this format:
[Tool Name] [Tier] [Usage Type] – [usage per day/week/month]
Usage Types:
[a] Heavy: Intensive coding, vibe coding, major refactors.
[b] Medium: Regular coding, debugging, smaller refactors.
[c] Low: Occasional quick fixes, snippets, reviews.
If you haven't hit limits yet, put NL (no limit).
codeclimberop19 hours ago
Gemini CLI (free tier) [a] - less than 1 h/day Claude Code (pro) [a] - 2 h/day Cursor (pro) [b] - 2 weeks/month
incomingpain18 hours ago
Gemini cli free tier, i can get about 2 million input tokens in about an hour before it dumps me down to flash.
Then i switch to local devstral for the rest of the day and naturally no limits in that sense.