Third-party cyber evaluations involving OpenAI models
openai.com·13h ago
Data Science
Towards AI·21d ago·5 min read
Codex Just Became Unlimited! Also GPT 5.6 Is Actually Really…
TL;DR
Codex has transitioned to an unlimited access model, allowing developers to utilize its capabilities without restrictions. This change aims to enhance productivity and creativity in coding tasks.
✦ Why It Matters
Engineers can immediately adopt Codex's unlimited model to enhance their coding workflows and explore new project ideas.
Key Takeaways
- Utilize the new weekly limit for extended coding sessions.
- Explore Sol, Terra, and Luna for cost-effective AI solutions.
- Monitor performance metrics on the Artificial Analysis Intelligence Index.
- Take advantage of OpenAI's free resets for Codex usage.
Full Summary
OpenAI has made significant changes to its Codex model by eliminating the restrictive five-hour usage cap, allowing users to engage in longer coding sessions under a weekly limit instead. This adjustment is particularly beneficial for developers who require extended periods for complex coding tasks.
Alongside this, OpenAI has introduced three new models: Sol, Terra, and Luna, which provide more affordable options compared to the previous GPT-5.5 model. Sol achieves a score of 59 on the Artificial Analysis Intelligence Index, closely trailing behind Claude Fable 5.
Terra offers a 55 score while halving the token rates of GPT-5.5. Luna, priced at $1 per million input tokens and $6 per million output tokens, rounds out the new offerings.
These changes not only enhance the usability of Codex but also present developers with cost-effective solutions for their AI needs.