OpenAI releases GPT-5.5 with agentic capabilities and doubled API pricing

OpenAI has released GPT-5.5, a new agent-based model designed to autonomously handle complex tasks across multiple tools, alongside the more capable GPT-5.5 Pro variant. The models are now available to paying ChatGPT and Codex users on Plus, Pro, Business, and Enterprise plans, with API access launching at double the cost of its predecessor.

On Terminal-Bench 2.0, GPT-5.5 scores 82.7 percent, outperforming Anthropic's Claude Opus 4.7 (69.4 percent) and Google's Gemini 3.1 Pro (68.5 percent).
The model achieves significant gains in long-context performance, reaching 74.0 percent on MRCR v2 at context lengths of 512K to 1M tokens.
API pricing for GPT-5.5 is set at $5 per million input tokens and $30 per million output tokens, exactly twice the rate of GPT-5.4.
Independent testing by Artificial Analysis notes that while GPT-5.5 takes the overall top spot in benchmarks, it shows notable weaknesses with hallucinations.

OpenAI positions GPT-5.5 as a "new class of intelligence" for real work, emphasizing its ability to plan, gather context, and complete longer workflows with less guidance.

Benchmark	Model	Score
BrowseComp	Gemini 3.1 Pro	85.9%
BrowseComp	GPT-5.5	84.4%
Terminal-Bench	GPT-5.5	82.7%
SWE-bench	Claude Opus 4.7	64.3%
SWE-bench	GPT-5.5	58.6%
FrontierMath	GPT-5.5 Pro	39.6%
FrontierMath	GPT-5.5	35.4%

Benchmarks