Artificial Analysis announced the launch of Intelligence Index v5, adding two new benchmarks (FrontierCode 2.0 and Terminal-Bench-Science) and upgrading three existing benchmarks, with 60% of the index based on private datasets for the first time. It also released Progress Index v1 to track model capability changes over time.
The Browser Company announced on October 22 that it will discontinue the Arc browser on December 22, guiding users to its AI-first browser Dia. The shutdown will occur in three stages, including the loss of AI features in Arc Search on November 23 and the complete termination of Arc on Mac and Windows on December 22, with account data deleted within 30 days.
Panay shared his views on the future of personal computing, saying ambient sensing devices will proactively understand user needs and reduce dependence on apps. He emphasized privacy and user control, citing Bee as an example, and predicted more sensor-driven devices while noting smartphones will remain central, with a smart layer connecting all devices.
Arcade.dev has launched Tool Recommendation, a feature that helps AI agents find suitable actions without repeatedly loading lengthy tool lists. The company says it can cut token use for tool discovery by up to 91%, reduce modeled input costs by 37%, and lower full-workflow token bills by 7.6%. The feature is available to Arcade Cloud customers starting October 15 at no extra charge.
Amazon-owned Bee unveiled its first products since the acquisition: the Bee Frontier ($74.99), a wearable that can be worn on the wrist or clipped to clothing, and the Bee Eclipse ($99.99), a credit-card-sized device that magnetically attaches to the back of a smartphone. The devices record everyday conversations and generate summaries, to-dos, and insights. The free tier provides 14 days of memory; the $11.99/month Bee Pro subscription unlocks unlimited summaries and integrations with Alexa+ and more.
On October 14, Artificial Analysis released the AA-Robotics v0.1 benchmark, testing frontier models such as GPT, Claude, and Gemini on controlling a real robotic arm to complete eight zero-shot tasks. GPT-6 Astra topped the ranking with 69.4 points, followed by GPT-6.1 Sol at 64.6 and Gemini 4 Argon at 55.3. On average, models took about half an hour per task, far slower than the human baseline.
Blueprints AI, founded in Santa Monica in 2025 by Al Drams and Coco Cao, uses AI to generate code-compliant, permit-ready construction blueprints from 2D sketches or PDFs in minutes. It has been selected for the 2026 Startup Battlefield 200 and will exhibit at TechCrunch Disrupt from October 13 to 15 in San Francisco, where it will unveil a more capable second-generation version of its software.
Alec Tarashansky, Anthropic technical team member and Claude Science development lead, announced he is joining AI healthcare company OpenEvidence. He holds a PhD in bioengineering from Stanford and worked at the Chan Zuckerberg Initiative for three years; he brought the Claude Science prototype to Anthropic and helped develop it for about seven months.
A four-person team in Atlanta, US, has developed ArtCraft, a free and open-source Adobe alternative suite, using AI-generated code. The project reverse-engineers Adobe software to remove cloud dependency and subscription fees. PhotoCraft targets Photoshop, while LightCraft targets Lightroom; all apps are still under development, with some in early alpha testing.
AI model evaluation platform Arena announced a $200 million Series B round at a $3.1 billion valuation, led by Lightspeed Venture Partners and Khosla Ventures, with Salesforce Ventures, a16z and others participating. The company reached $100 million in annualized revenue in June; it had previously raised a $150 million Series A at a $1.7 billion valuation in January.