AI News for July 10, 2026 — 7min.ai

[NEWS] GPT-5.6 goes live globally; Altman touts 54% efficiency gain, prices Sol at a third of Fable 5: OpenAI's GPT-5.6 family, Sol, Terra, and Luna, is now publicly available worldwide, ending the two-week limited preview that had restricted access to government-vetted partners.
[TOOLS] OpenAI unveils ChatGPT Work, an agent that runs full workflows for hours: OpenAI launched ChatGPT Work, an agent built on GPT-5.6 and Codex that stays on a project for hours and turns a goal into finished documents, spreadsheets, presentations, and web apps.
[NEWS] Meta launches Muse Spark 1.1, its first paid model, at cut-rate prices: Meta released Muse Spark 1.1, a multimodal reasoning model for agentic coding and computer use, and opened its first developer API.
[NEWS] OpenAI's No. 2 Fidji Simo steps down over chronic illness: Fidji Simo, OpenAI's CEO of applications and its de facto No.
[NEWS] OpenAI pushes back on Microsoft-MAI swap reports, calls GPT-5.6 Copilot's 'preferred model': OpenAI announced GPT-5.6 will be the 'preferred model' powering Microsoft 365 Copilot across Word, Excel, PowerPoint, Chat, and Cowork, moving to quell speculation that the two companies are drifting apart.
[NEWS] OpenAI shuts down ChatGPT Atlas browser, redistributes its features: OpenAI is sunsetting ChatGPT Atlas, the AI browser it launched in October, targeting August 9 for deprecation.
[NEWS] Anthropic hits $1.2 trillion secondary valuation as shares vanish: Anthropic soared to a $1.2 trillion valuation on secondary markets, a 550% YoY jump, but almost no one will sell.
[RESEARCH] OpenAI's AI beats every human at AtCoder programming world finals: At the AtCoder World Tour Finals 2026 in Japan, an OpenAI system solved all five problems in the Algorithm Division and finished first by a wide margin against the world's top competitive programmers.
[RESEARCH] GPT-5.6 Sol nearly matches Fable 5 on benchmarks at a third of the cost: GPT-5.6 Sol scored 59 on Artificial Analysis's Intelligence Index, one point behind Claude Fable 5, while costing about $1.04 per task versus roughly $3 for Fable.
[RESEARCH] Google Research debuts SensorFM, a foundation model for wearable health data: Google Research introduced SensorFM, a foundation model aimed at building general intelligence and a natural-language interface over data from wearable health sensors.

About 7min.ai · Subscribe to newsletter