GPT-6 Astra
OpenAI's GPT-6 Astra reaches the Critical cybersecurity capability threshold, with computer use, a 1.05M token context, and reasoning up to max effort.
Model Overview
Capabilities, design details, and architectural traits
GPT-6 Astra - OpenAI's flagship model
GPT-6 Astra is OpenAI's flagship model for complex reasoning and coding. It is the first model OpenAI has designated at the Critical cybersecurity capability threshold under its Preparedness Framework, meaning that with the right tools and access it can find previously unknown security flaws and develop exploits for them without a person guiding each step.
What sets GPT-6 Astra apart
| Trait | Detail |
|---|---|
| First Critical-cyber designation | First OpenAI model to meet the Critical cybersecurity capability threshold under the Preparedness Framework, requiring stronger safeguards during development and before release |
| Zero-day discovery | During evaluation it discovered and used two zero-day V8 vulnerabilities as part of an exploit chain, which OpenAI is disclosing to maintainers |
| Reasoning without an off switch | Reasoning effort runs from low to max with no "none" setting, unlike the GPT-5.6 family |
| Scale | 1.05M token context window and 128K token max output |
| Built-in tools | Functions, Web search, File search, and Computer use |
Professional work and efficiency
Astra pairs computer-use advances with targeted training for professional environments. It carries out multistep workflows and produces polished documents, spreadsheets, and presentations.
Guarded cyber access
Because of its critical cyber capabilities, Astra's most advanced cybersecurity work is not in the default production configuration. It is initially limited to a group of testers, with defensive access expanding through Daybreak Blue.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | OpenAI | Anthropic | Anthropic |
| Release Date | September 3, 2026 | September 22, 2026 | September 28, 2026 |
| Knowledge Cutoff | Apr 2026 | - | Jun 2026 |
| Context & Limits | |||
| Context Window | 1.1M Best Context Window | 1M | 1M |
| Pricing (per 1M tokens) | |||
| Input Pricing | $10 | $4 | $2 Best Input Pricing |
| Output Pricing | $50 | $20 | $10 Best Output Pricing |
| Modalities | |||
| Inputs | textimagefile | textimagefile | textimagefile |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 52.7 | 57.6 Best Intelligence Index | 56.0 |
| Coding Index | 76.9 | - | - |
| Agentic Index | 51.0 | - | - |
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
SciCode Benchmark
Scientific coding and mathematical modeling.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.
Explore more from OpenAI
Other models by OpenAI
Top AI Models
Leading alternatives by intelligence score
