2026
10,546 articles · 132 days · 10 categories · avg 80 per day
Anthropic releases Claude Haiku 5.5: 43 points on the AA index and a price of $0.10 per million tokens
Claude Haiku 5.5 scores 43 on the Artificial Analysis Intelligence Index, drops to $0.10/$0.50 per million tokens and gets 1M context, but uses three times the…
Three OpenAI Researchers Deny Leaks and Warn of a Cultural Crisis
Jasmine Wang, Tomek Korbak, and Mikita Balesni have rejected leak accusations and warned of the chilling effect their firing will have inside a company where staff…
Odyssey opens preview of its 14-billion-parameter Odyssey-3 world model
Odyssey has released Odyssey-3, a 14-billion-parameter model. It generates interactive worlds in real time, leads physics benchmarks, and can control robot arms…
Microsoft Prepares to Install Gigawatts of Its Own Chips in Data Centers
Claude Can Now Run Up to 1000 Subagents and Find 66 of 70 Bugs
AI buildout has borrowed hundreds of billions in the bond market and is crowding out other issuers
Maia 300: How Microsoft Is Cutting Its Dependence on Nvidia GPUs With Its Own AI Chip
Nvidia in talks to acquire or invest more in Reflection AI
Nvidia is in early talks to acquire Reflection AI or invest more in the US open-weight model developer, the Financial Times reports. Terms have not been disclosed.
2.14 Million Banned Accounts and a 3.3% Appeal Success Rate: How Claude's Anti-Fraud and Ban Protection System Works
According to Anthropic's official Transparency Hub report, the company blocked 2.14 million accounts in 2025 (a 110% jump in the second half of the year to 1.45…
Anthropic admits Claude took more unauthorized actions, including on US government websites
DistroKid Pulls Tracks Without Warning After UMG Lawsuit Over "AI Slop"
Anthropic gives METR employee-level access to evaluate its systems' safety
Anthropic cuts internet access for all internal model tests after incidents
DynaFlow teaches a lead agent to write its own workflows, beats the baseline by 10.9 points
NUS researchers trained a lead agent on Qwen3.5-27B to write its own programs for sub-agents. Result: 88.08% accuracy across six benchmarks, plus 10.90 points over…
Alibaba DAMO Trains Robots on 30,000 Hours of Video Without Action Labels
RynnVLA-Latent from Alibaba DAMO Academy pretrains on 30,100 hours of video with no action annotations. The 4-billion-parameter model scored 98.8% on LIBERO and…
AMBIT trains one model to generate and edit 3D sound from text
NoizAI's team has unveiled AMBIT: a text command turns into a ScenePlan with sources and 3D trajectories, and a single Transfusion model renders Ambisonics and…
Harvey LAB-AA v1.1: Grok 4.7 Leads With 9.4% After Hallucination Check
Anthropic's agents submitted 20 visa applications on a US State Department website
Meta releases Muse Wonder and Wonder Full: intelligence index rises from 48 to 56
Cloudflare Buys Deno and Will Wind Down the Runtime in a Year
Moonshot releases Kimi K3.1 Pro: a 3T open-weight model with a 1M-token context window
Moonshot unveils Kimi K3.1 Pro, which scores 65.7% on Terminal-Bench 4.0 and beats GPT-6 Astra. Weights are due by November 4. Context window: 1M tokens.
Artificial Analysis launches AA-3D: Claude Sonnet 5.5 leads with 1194 Elo
A new benchmark asks agents to build an interactive 3D scene in three.js. Claude Sonnet 5.5 takes first place (1194 Elo), but the top four are statistically…
VentureBeat: 79% of Companies Leave GPUs Idle at Least Half the Time
Moonshot AI Unveils Kimi K3.1 Pro, the First Pro Model in the Lineup, After the Kibitz Test
Inception Releases Mercury 3: a Diffusion LLM 8x Faster Than GPT-6 Luna
Matthew Green Puts 15% Odds on AI Eroding Trust in Public-Key Cryptography
Russia Already Has 18 AI Code Agents, but 'Russian' Means Three Different Things
RuBench's census counted 18 Russian AI code agents (10 for IDE/CLI and 8 builders): from GigaCode, SourceCraft, and Kodik to T-Code, Koda, and 1C:Naparnike. We…
Oracle Cuts Tasks From Days to Minutes With ChatGPT Work and Codex
Oracle uses ChatGPT Work and Codex in recruiting, engineering and operations. According to OpenAI, tasks that once took days now take minutes.
Alibaba Qwen Names Three Pillars of Its Personal Agent: Model, Memory and Ecosystem
At CCF 2026, Qwen algorithm lead Jiang Guanjun explained how Alibaba is building a personal agent: the model, memory and ecosystem work only in combination.
Eight Security Assumptions AI Agents Are Breaking: From Two Weeks to Ten Hours
Devin Gets 30–40% Cheaper as Fusion Tops FrontierCode 1.1 With a 68.8 Score
Google teaches EmbeddingGemma 2 to understand text, images, video, and audio
