OpenAI’s GPT-5.5 Launches with Improved Efficiency and Higher Costs
- GPT-5.5 is now available for Plus, Pro, Business, and Enterprise users in ChatGPT and Codex, with API access soon at $5 per million input tokens and $30 per million output tokens.
- The model achieves a score of 82.7% on Terminal-Bench 2.0, surpassing Claude Opus at 69.4% and Gemini Pro at 68.5%.
- GPT-5.5 uses fewer tokens than GPT-5.4 for the same tasks, improving efficiency despite higher pricing.
- On GDPval benchmark testing across various occupations, GPT-5.5 matches or exceeds industry professionals in nearly all comparisons.
OpenAI’s latest model, GPT-5.5, offers enhanced performance in complex workflows while maintaining token efficiency over its predecessor GPT-5.4 despite increased costs per token usage.
GPT-5.5’s higher benchmark scores illustrate its advanced capabilities in agentic coding environments like Codex, marking significant progress in AI-driven computer use.Source
