Particle.news

OpenAI Releases GPT-6.2 Astra and Sol

An enterprise-focused launch lowering token costs with claimed benchmark gains over rival models.

Overview

  • OpenAI announced two new models, GPT-6.2 Astra as a professional flagship and GPT-6.2 Sol as a lower-cost general model, and made them available to ChatGPT Work, Codex and API customers on paid tiers while not adding them to the consumer chat product.
  • OpenAI says Astra improves multi-step agent workflows, scientific reasoning, coding and design tasks and that it outperforms prior OpenAI generations and a leading competitor on named tests such as Agents' Last Exam, Terminal-Bench-Science 0.1, Vals Science Index and AA-3D.
  • The company published API prices with GPT-6.2 Astra at $8 per million input tokens, $0.4 per million cached input tokens, and $40 per million output tokens, and GPT-6.2 Sol at $2 input, $0.1 cached input, and $10 output per million tokens.
  • Access is gated: Pro, Plus, Enterprise and Edu users on ChatGPT Work and Codex can use the models, Pro 500 gets Ultrafast mode for both models, Pro 200 gets Ultrafast for Sol, and Enterprise/Edu admins must enable the features for their workspaces.
  • OpenAI frames the release as a cost-and-performance play that could accelerate enterprise and developer adoption if independent tests confirm the claims, and observers note the timing may signal competitive pressure on rivals such as Anthropic.