Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until w…

SemiAnalysis Twitter · SemiAnalysis (@SemiAnalysis_) · 2026-06-10

SemiAnalysis reports empirical testing of Anthropic and OpenAI subscription plans by running long-horizon coding tasks to exhaustion, teasing findings that challenge the assumption that a $200/month plan delivers roughly $2,000/month in API-equivalent token value.

Open original ↗

Appears in

SemiAnalysis: AI Subscriptions Are 40–70x Cheaper Than API for Heavy Users; OpenAI Eyes Deep Price Cuts

Extraction

Topics: ai-subscription-pricingai-economicsanthropicopenaitoken-limits

Claims

SemiAnalysis empirically tested AI subscription plans by purchasing each tier from both Anthropic and OpenAI and running long-horizon coding tasks until weekly usage limits were hit.
The widely cited assumption that a $200/month AI subscription plan provides approximately $2,000/month worth of tokens at API pricing may not reflect real-world observed limits.

Key quotes

Recently, we purchased one of each Anthropic/OpenAI subscription plan and randomly ran long horizon coding tasks until we exhausted the weekly limit. It's widely believed that a $200/month plan maxes out at ~$2000/month worth of tokens (assuming API pricing). However, we found