OpenAI launches GPT-6 Astra, a computer-use model with a 1.05M-token context window and staged enterprise/API access

Decision brief

What changed

OpenAI released GPT-6 Astra as a new flagship computer-use model designed for multi-step tasks across tools, with a 1,050,000-token context window. Initial access is gated through OpenAI programs such as Daybreak and Trusted Access, with broader staged availability planned, and API pricing was published at $10 per 1 million input tokens and $50 per 1 million output tokens.

Why it matters

This creates a near-term decision point on whether to evaluate a higher-autonomy model that is explicitly positioned for workflow execution rather than chat, but only if your organization can qualify for staged access or wait for broader rollout. The published token pricing and large context window give leaders an early basis for budgeting pilots and comparing Astra against existing vendors for engineering, operations, or cybersecurity use cases. Because access is gated and protections are emphasized, adoption decisions should balance potential productivity gains against deployment constraints, governance requirements, and uncertain real-world performance outside OpenAI’s cited benchmarks.

Affected roles

CEO CFO COO CTO CISO

Evidence

The coverage is consistent across four outlets that OpenAI launched GPT-6 Astra, set API pricing at $10 per 1M input tokens and $50 per 1M output tokens, and is gating access through Daybreak/Trusted Access before broader expansion. The New Stack, Trending Topics EU, and MarkTechPost independently align on the staged rollout and pricing, while Latent Space adds reported benchmark and workflow automation claims that appear based on launch materials rather than independent validation.

What remains uncertain

Open questions include how quickly access will expand beyond selected programs, what usage restrictions or qualification thresholds will apply, and whether benchmark-level performance translates into reliable production outcomes for enterprise workflows. Claims about end-to-end AI engineering capability, hourly operating cost, and competitive advantage over alternatives are not independently verified in the provided coverage and should be treated as assumptions until tested in pilot deployments.

Monitor next

Watch for OpenAI’s next concrete access expansion announcement—especially general API availability and any published enterprise case studies or usage limits for Astra.

Analytical support, not advice — assumptions and open questions stated above.

Source coverage