Leaked OpenAI Models and Work Mode Details
Ben's Bites · rss · 2026-07-14
Ben’s Bites summarized a batch of new OpenAI updates, focusing on the so-called GPT-5.6 model series and the ChatGPT/Codex integration.
Key details include:
- The ChatGPT macOS App and Codex App have merged, introducing a new ChatGPT Work mode
- A new ChatGPT Sites plugin allows for website hosting with "Login with ChatGPT" support
- GPT-5.6 is split into three models: Luna / Terra / Sol, each with multiple thinking levels, plus a new Ultra mode that triggers more subagents
- Higher thinking levels consume credits faster; the author recommends using lighter modes for daily tasks
- Author's impressions: Sol leans toward UI and writing, Terra feels like a more controllable upgrade to 5.5, and Luna acts like a mini model but avoids failing on clear tasks
- OpenAI reportedly reset usage limits multiple times to fix bugs from the app merger and temporarily removed the 5-hour limit
- The new version performs better in Computer Use, capable of viewing screens, clicking buttons, and operating across apps
More from Models
- Same Echo Maze prompt, three frontier models: all passed visually but shipped the same hidden bug — eyishazyer · 2026-09-11
- Benchmark scores drop from 89% to 19% on new evals — how benchmaxxing breaks leaderboard trust — airesearch12 · 2026-09-11
- ChatGPT tells user their question is too hard and to 'accept dumber answers' — phido3000 · 2026-09-11
- Claude is no longer available for minors as Anthropic rolls out age assurance — Muhammad523 · 2026-09-11
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11