OpenAI Launches GPT-5.5: A Breakthrough AI Model for Coding, Science, and Autonomous Tasks

OpenAI Unveils GPT-5.5: A Leap Forward in AI Capabilities

On Thursday, OpenAI announced the release of GPT-5.5, its most capable AI system to date. The model is engineered to enhance the performance of the Codex coding agent, enabling it to tackle a broader range of general digital work tasks with greater efficiency and precision.

GPT-5.5 represents a significant advancement in autonomous and agentic capabilities. According to OpenAI’s blog post, the system is designed to complete complex, multi-step tasks on a computer without human guidance, marking a critical step toward fully autonomous AI workflows.

Superior Performance on Key Benchmarks

GPT-5.5 has achieved top scores on several industry-standard benchmarks, demonstrating its superiority over previous models and competitors:

• Terminal-Bench 2.0 (tests complex command-line workflows):
  • GPT-5.5: 82.7%
  • GPT-5.4: 75.1%
  • Anthropic’s Opus 4.7: 69.4%
  • Google’s Gemini 3.1 Pro: 68.5%
• OSWorld-Verified (measures independent computer operation):
  • GPT-5.5: 78.7%
  • GPT-5.4: 75.0%
  • Anthropic’s Opus 4.7: 78%

While Anthropic’s newest frontier model, Mythos, has been completed, it has not yet been publicly released.

Enhanced Coding and Scientific Capabilities

OpenAI reports a surge in users for its Codex coding agent, with approximately 4 million developers using the tool weekly. During a press call, OpenAI CEO Greg Brockman highlighted that GPT-5.5 will enable Codex to generate polished, production-ready code and execute coding projects with the judgment of a senior software engineer.

On the SWE-Bench Pro benchmark, which evaluates real-world GitHub issue resolution, GPT-5.5 resolves 58.6% of tasks end-to-end in a single pass. Early testers noted that the model demonstrates a deeper understanding of software system architecture, including the ability to identify why failures occur, pinpoint necessary fixes, and assess the broader impact on the codebase.

Rapid AI Development and Market Competition

The release of GPT-5.5 follows closely behind GPT-5.4, underscoring OpenAI’s commitment to rapid innovation. The company faces increasing pressure from rival Anthropic, particularly in the realm of AI-powered coding assistants, which have become a cornerstone of AI’s impact on business operations. OpenAI’s accelerated release cycle is partly attributed to AI’s role in automating the development of subsequent AI systems.

Speculation suggests that GPT-5.5 may feature up to 10 trillion parameters, though Brockman declined to confirm this figure.

Availability and Subscription Tiers

GPT-5.5 is now available to users across OpenAI’s subscription tiers, including Plus, Pro, Business, and Enterprise, within both ChatGPT and Codex. Additionally, OpenAI offers GPT-5.5 Pro, a higher-accuracy version, exclusively to Pro, Business, and Enterprise users.

"GPT-5.5 represents a step toward AI systems that can complete complex, multi-step tasks on a computer without human guidance."