Anthropic has disclosed its S-1 filing to the U.S. SEC, planning to list on Nasdaq under ticker ANT, targeting a listing in the week of 2026, month 11, day 16. The filing shows H1 2026 revenue of $162.86 hundred million, up about 13x year over year, a net loss of $807.48 hundred million, and future compute procurement obligations of about $4170 hundred million. Market valuation expectations have been cut from $2 trillion to roughly $1.5 trillion.
According to Ebrun, Kuaishou announced on the Hong Kong Stock Exchange that its video generation large model Kling AI has completed the closing of this round of capital increase, with all investors having completed payment. The actual amount received was approximately RMB 198.02 hundred million (about 29.05 hundred million USD), with a post-investment valuation of approximately 180 hundred million USD. The actual amount received was approximately RMB 6.45 hundred million lower than the previously announced subscription cap of RMB 204.47 hundred million, with the shortfall resulting from two originally planned investors withdrawing from subscription.
On the evening of 10 month 30, Kuaishou announced on the Hong Kong Stock Exchange that its video generation large model Kling AI has completed this round of capital increase, with investors paying a total of approximately RMB 198 hundred million, exceeding US$29 hundred million. Two institutions terminated their subscriptions due to failure to pay on time, and part of the shares held by Dakanglanwu was transferred to Kingdee Software (China) and other transferees. In mid-10, Kling AI released its new-generation model Kling4.0, with over 1 hundred million global users and serving more than 5 thousand enterprises.
The 2026 World Internet Conference Wuzhen Summit opened on 11 3 under the theme "Open Source, Open Collaboration — Jointly Building a Community with a Shared Future in Cyberspace", featuring 27 sub-forums. On the eve of the summit, DeepSeek announced the full open-sourcing of its internal inference framework along with codec components. The "Light of the Internet" Expo brought together 671 institutions from 54 countries and regions, showcasing over 1000 new technologies, including more than 30 embodied intelligence robots.
ModelBest has released and open-sourced the first OCR series model in its MiniCPM family, including the 1.3B-parameter parsing model MiniCPM-OCR 1.0 and the 2B-parameter verification decision model MiniCPM-OCR Jev. The former scored 93.68 on Real5-OmniDocBench, with RTX 5090 measured throughput of 1.217 pages/second; the latter outputs 0-1 trustworthiness scores, with JevOCRBench overall AUROC of 93.6. All weights are open, with adaptations for inference frameworks such as vLLM and SGLang and 12 hardware platforms completed on day one.
Kunlun Wanwei disclosed its 2026 Q3 report on the evening of 10 29: revenue for the first three quarters was RMB 89.37 billion, up 53.96% YoY, with net profit attributable to shareholders of RMB 260.94 million, turning from loss to profit YoY. Among this, Skywork AI business revenue was RMB 43.21 billion, up 168.29% YoY, and overseas revenue was RMB 87.06 billion. During the reporting period, the company iterated its music, game world models and super agent products, while its subsidiary also launched the world action model Riemann-1.0 in the embodied intelligence space.
01.AI has reached a partnership with Hangzhou City and Linping District, with state-owned capital at both city and district levels jointly investing a combined several hundred million yuan. The company's headquarters will be located in Linping District, Hangzhou, which will serve as the foundation for building an industrial AI transformation base. The partnership comes at a critical stage as the large-model company founded by Kai-Fu Lee accelerates commercialization and advances its Pre-IPO strategic positioning; it had previously set targets of 15 hundred million yuan in order value by 2026 and an IPO by 2027.
At the 2026 Zhiyuan Innovation Day, Zhiyuan Co-founder Yao Maoqing unveiled the new-generation embodied foundation model GO-3 Preview. The company says the model has topped the three major evaluation leaderboards—RoboDojo, RoboTwin, and RoboColiseum—and plans to fully open-source it by the end of 2026. The model adopts Zhiyuan's proprietary embodied-native architecture WRAM (World Reasoning Action Model), which unifies the modeling of understanding, prediction, and action, and combines the Genie Evolver post-training system with the Genie Master agent framework to form an end-to-end learning system spanning pre-training, post-training, and post-deployment.
Microsoft AI released the image editing model MAI Image 2.7 as a preview on 10 27, 2026. Officially, it ranks No. 2 on the Artificial Analysis Image Editing leaderboard and No. 3 on the Arena Edit leaderboard, and outperforms the 2.6 version across all 10 use case categories and 7 edit operations covered by the evaluation. The update focuses on control in complex editing, covering style adjustments, relighting, text modification, and asset repair.
According to TMTPost, Tencent Hunyuan's simultaneous interpretation large model Hy-SimulMT was released and open-sourced on 10 26. The model emphasizes low latency and high accuracy, supports real-time Chinese-English speech-to-speech translation, and features real-time voice cloning to reproduce the speaker's voice and tone. It debuted on Tencent's Hy Translation APP with a free experience.
Baidu has released and open-sourced Ernie Unlimited OCR 2, which processes ultra-long documents through Sparse Reference Sliding Window Attention (SR-SWA) and the Unlimited-Horizon Decoding Paradigm. According to the technical report, in hundred-page single-sequence tests, peak KV cache dropped from 47GB to 1.77GB, a reduction of about 96%; 16-way concurrent throughput increased from 360 to 1315 tokens per second. A real-world test on the 884-page The Lord of the Rings completed parsing in about 43 minutes, with MPDocBench and MarathonDocBench overall scores of 84.59 and 77.06 respectively, both the highest among evaluated models.