OpenAI's new GPT-5.5 arrives with advanced agentic coding and computer use

OpenAI has officially launched its newest flagship frontier model, GPT-5.5, marking a significant leap in agentic capabilities.

OpenAI today launches its flagship frontier model, GPT-5.5. This new model performs significantly better in agentic coding, computer use, knowledge work, and early scientific research when compared to GPT-5.4 and other competing models.

According to OpenAI, GPT-5.5 is designed to handle a wide range of tasks, including writing and debugging code, online research, data analysis, document creation, and more. Instead of requiring users to prompt every step manually, this new model can plan, use tools, verify its own work, and continue operating to achieve the goal. Despite the jump in capability, OpenAI claims GPT-5.5 matches GPT-5.4’s per-token latency in real-world serving and also uses fewer tokens to complete the same Codex tasks.

The new GPT-5.5 model tops Artificial Analysis’s Coding Agent Index at roughly half the cost of competing frontier coding models. On Terminal-Bench 2.0, which measures performance on complex command-line workflows involving planning, iteration, and tool coordination, GPT-5.5 scored 82.7%. On SWE-Bench Pro, which tests real-world GitHub issue resolution, it reached 58.6%.

On GDPval, which measures how well agents can complete well-specified knowledge work across 44 occupations, GPT-5.5 scored 84.9%. On OSWorld-Verified, which tests autonomous computer use in real environments, it reached 78.7%. On Tau2-bench Telecom, which evaluates complex customer-service workflows, it scored 98.0% without prompt tuning.

Find the full benchmarks comparison below:

GPT-55 benchmarks

The GPT‑5.5 model is now available to ChatGPT Plus, Pro, Business, and Enterprise users in ChatGPT and Codex. The GPT‑5.5 Pro model is now rolling out to ChatGPT Pro, Business, and Enterprise users in ChatGPT. In Codex, GPT‑5.5 will also be available in Fast mode (1.5x faster) for 2.5x the cost.

GPT‑5.5 will be available for developers via API in the Responses and Chat Completions APIs at $5 per 1M input tokens and $30 per 1M output tokens, with a 1 million context window. As always, developers can get a 50% discount when using Batch and Flex processing. GPT‑5.5 Pro will be available for developers via API in the Responses and Chat Completions APIs at $30 per 1M input tokens and $180 per 1M output tokens. OpenAI highlighted that GPT‑5.5 will deliver better results with fewer tokens when compared to GPT-5.4.