bibibi.ccLoading…
Connect scattered updates into a story you can understand.
InclusionAI releases Ling-3.1-flash, with a clear increase in parameter scale and per-token activation scale compared with the previous generation. The official demonstration showed it completing long-duration programming tasks, with plans to make context and an open-source model available later.
The PANews Telegram channel relayed a report stating that OpenAI DevDay 2026 released the Dots Personal Agent,GPT-6.1 Sol, and 500 USD Pro membership, among other product and pricing information. This information has not yet been corroborated by a primary OpenAI announcement.
Anthropic releases Claude Sonnet 5.5，API Pricing remains unchanged from the previous generation; the company says speed and task efficiency have improved, and it is now available on multiple platforms.
Anthropic released Claude Sonnet 5.5, saying it outperforms its predecessor in programming and creating documents, slides, and spreadsheets, with significantly improved cybersecurity capabilities.
According to Beating AI flash news, NaiveAI, founded by Tsinghua University associate professor Dai Jifeng, released Naive-N0.5-Flash and made the model weights available under the MIT protocol. The model was adapted from Xiaomi MiMo-V2.5 Base.
Vercel AI Gateway Anonymous model Pixel Canary launches and is available for free for a limited time before its identity is disclosed; Cline has integrated it. Vercel says the model’s baseline success rate in its Next.js evaluation is 90%。
Meituan releases LongCat-2.5-Preview, adding native multimodal capabilities and integration with the official API, supporting long-horizon Agent tasks;benchmark data has not yet been disclosed.
OpenRouter launches Batch API, offering pricing for input and output of most models for non-immediate tasks, with results guaranteed to be returned within Token pricing, with results guaranteed to be returned within 24 hours.
A PANews article states that the AI model Jev, focused on decision-making tasks, reached 24 hours after release Vercel approximately 13% paid teams, but the material provides neither a statistical basis nor independent corroboration.