OpenAI has introduced a new $100 ChatGPT Pro subscription tier, expanding its offerings. This plan joins existing individual options including free access, an $8 monthly Go plan, a $20 monthly Plus plan, and a $200 monthly Pro plan. OpenAI also provides specialized plans for organizations, such as Edu, Business, and Enterprise.
The main benefit of the new $100 Pro plan is a five-fold increase in usage limits for Codex, OpenAI’s AI application designed for coding, compared to the $20 ChatGPT Plus plan. OpenAI CEO Sam Altman stated the launch was due to high demand for Codex.
Simultaneously, OpenAI announced adjustments to the $20 ChatGPT Plus plan. The company indicated it is “rebalancing Codex usage” in Plus to support more frequent, shorter sessions throughout the week, implying a recalibration of daily usage allowances for Plus subscribers.
The $20 ChatGPT Plus plan offers baseline usage limits, including 33–168 local messages for GPT-5.4 and 45–225 local messages with 10–60 cloud tasks for GPT-5.3-Codex over five hours. The new $100 ChatGPT Pro plan significantly expands these, providing roughly five times the capacity. For example, GPT-5.4 local messages increase to 223–1,120, alongside substantial boosts for other models, cloud tasks, and code reviews. It also provides access to GPT-5.3-Codex-Spark. The $200 Pro plan doubles the $100 tier’s capacity.
This strategic move by OpenAI occurs amidst heightened competition with rival Anthropic, which recently reported annualized run-rate revenue exceeding $30 billion, surpassing OpenAI’s last reported figures. Anthropic’s growth was driven by its enterprise coding products, Claude Code and Claude Cowork. Earlier this month, Anthropic restricted Claude subscriptions from powering third-party agentic AI harnesses like OpenClaw. Users must now pay via Anthropic’s API or additional usage credits for such integrations.
The timing of OpenAI’s new tier, providing capacity similar to what Anthropic recently restricted, suggests a direct response. Peter Steinberger, OpenClaw’s creator and hired by OpenAI in February 2026 to lead its personal agent strategy, has criticized Anthropic’s limitations, positioning OpenAI’s Codex models with fewer restrictions. This move appears aimed at attracting professional developers and the OpenClaw community.
Source: Original
