Replit Agent switches from fixed monthly quota to opaque 'effort-based' billing
Replit's $25/mo Core plan moved from a deterministic monthly agent quota to a metered 'effort units' system. Users now pay per task with no upfront indication of cost. CEO Amjad Masad publicly acknowledged in posts that 'every Cursor-like company is paying more to Anthropic/OpenAI than they make.'
$25 / mo + opaque effort units · 2-4x cost per task
Replit did not respond inside the 30-day window. This receipt now reads as uncontested in the public record.
The transition happened in two phases — first an opt-in beta, then a forced migration in November.
The transition happened in two phases — first an opt-in beta, then a forced migration in November. Users report tasks that previously cost ~5% of monthly quota now cost 20-30% with no obvious driver of the increase. The 'effort' metric is undocumented.
Open-source Claude Code / Cursor replacement that lives in the terminal. Pick when you want zero vendor lock-in and any LLM backend.
Permissive (MIT) open-weight model targeting ChatGPT/Claude-class chat and agentic coding without API lock-in - the strongest open option for a self-hosted stack.
ChatGPT's Code Interpreter, but local and against any model.
Near-frontier open-weight model to back a self-hosted ChatGPT/Claude replacement or an agentic coding stack.