{"slug": "openai-might-be-preparing-gpt-5-6-for-next-week-s-release", "title": "OpenAI might be preparing GPT-5.6 for next week's release", "summary": "OpenAI has moved GPT-5.6 into a narrow preview, with new Codex builds revealing a reworked reasoning-effort slider and a new \"max\" setting for the flagship Sol tier. The release, rumored for next week, comes as Anthropic's Fable 5 moves to usage-based credits on July 7, potentially driving users to OpenAI. Broad access remains tied to US government cybersecurity review approvals.", "body_md": "OpenAI has moved **GPT-5.6** into a narrow preview, and fresh signals in the Codex app hint at how the company plans to surface it once the gate lifts. The family, unveiled on June 26, splits into three tiers: Sol as the flagship, Terra as a mid-cost option, and Luna as the fastest and cheapest. For now, the trio reaches only vetted partners through Codex and the API, with no place in ChatGPT.\n\nWhat stands out in recent Codex builds is a **reworked reasoning-effort control**, rendered as a slider rather than preset buttons. This tracks with a confirmed part of GPT-5.6: a new \"max\" setting that gives Sol more room to work through long problems, plus a subagent-driven \"ultra\" mode for heavier jobs. A slider would hand developers one control to trade speed against depth. The layout resembles the reasoning selector already on Anthropic's Claude Code desktop client, though it could shift before anything ships. References to real-time voice, present in earlier builds, also appear to have been removed from the current app, with no clarity on whether the capability will still be developed for Codex.\n\nThe developers and enterprises leaning on Codex for agentic coding stand to feel this first, and the timing is pointed. Anthropic's Fable 5, restored globally on July 1, **will no longer be bundled into subscription plans on July 7** and will move to usage-based credits, prompting some users to weigh alternatives just as OpenAI courts them. The Sol, Terra, and Luna labels mark a shift toward durable capability tiers that each advance on their own schedule, a rethink of how the company names and prices its lineup. Broad access is **rumored for the same window** but stays tied to voluntary US government review under a recent cybersecurity order, so approvals, not a fixed calendar, will decide when GPT-5.6 opens to everyone.", "url": "https://wpnews.pro/news/openai-might-be-preparing-gpt-5-6-for-next-week-s-release", "canonical_source": "https://www.testingcatalog.com/openai-might-be-preparing-gpt-5-6-for-next-weeks-release/", "published_at": "2026-07-04 14:08:03+00:00", "updated_at": "2026-07-04 14:22:10.402283+00:00", "lang": "en", "topics": ["large-language-models", "ai-products", "ai-tools", "ai-infrastructure", "ai-agents"], "entities": ["OpenAI", "GPT-5.6", "Codex", "Anthropic", "Fable 5", "Sol", "Terra", "Luna"], "alternates": {"html": "https://wpnews.pro/news/openai-might-be-preparing-gpt-5-6-for-next-week-s-release", "markdown": "https://wpnews.pro/news/openai-might-be-preparing-gpt-5-6-for-next-week-s-release.md", "text": "https://wpnews.pro/news/openai-might-be-preparing-gpt-5-6-for-next-week-s-release.txt", "jsonld": "https://wpnews.pro/news/openai-might-be-preparing-gpt-5-6-for-next-week-s-release.jsonld"}}