z.ai debuts faster, cheaper GLM-5 Turbo model for agents and ‘claws’ — but it’s not open-source
Carl Franzen March 16, 2026 Credit: VentureBeat made with Google Gemini 3.1 Pro Image Chinese AI startup Z.ai, known for…
model
Databricks built a RAG agent it says can handle every kind of enterprise search
Sean Michael Kerner March 5, 2026 Credit: Image generated by VentureBeat with FLUX-2-ProMost enterprise RAG pipelines are optimized for one…
Google releases Gemini 3.1 Flash Lite at 1/8th the cost of Pro
FeaturedCarl Franzen March 3, 2026 Credit: VentureBeat made with Google Gemini 3.1 Pro ImageGoogle’s newest AI model is here: Gemini…
Google reveals Nano Banana 2 AI image model, coming to Gemini today
Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more…
Spanish ‘soonicorn’ Multiverse Computing releases free compressed AI model
Large language models have a problem: they are large. Multiverse Computing, a Spanish startup, is addressing this issue with compressed…
Google’s new Gemini Pro model has record benchmark scores — again
In Brief Posted: 4:55 PM PST · February 19, 2026 Image Credits:Jagmeet Singh / TechCrunch Lucas Ropek Google’s new Gemini…
Google launches Gemini 3.1 Pro, retaking AI crown with 2X+ reasoning performance boost
FeaturedCarl Franzen February 19, 2026 Credit: VentureBeat made with Google Gemini 3 Pro Image on fal.aiLate last year, Google briefly…
Gemini 3.1 Pro
Gemini 3.1 Pro: A smarter model for your most complex tasks Feb 19, 2026 · Share x.com Facebook LinkedIn Mail…
Rapidata emerges to shorten AI model development cycles from months to days with near real-time RLHF
FeaturedCarl Franzen February 19, 2026 Credit: VentureBeat made with Flux 2 Pro on fal.aiDespite growing chatter about a future when…
Alibaba’s Qwen 3.5 397B-A17 beats its larger trillion-parameter model — at a fraction of the cost
Sam Witteveen February 18, 2026 Credit: VentureBeat made with GPT-Image-1.5Alibaba dropped Qwen3.5 earlier this week, timed to coincide with the…
Claude Sonnet 4.6
ProductIntroducing Claude Sonnet 4.6Feb 17, 2026Claude Sonnet 4.6 is our most capable Sonnet model yet. It’s a full upgrade of…
OpenAI removes access to sycophancy-prone GPT-4o model
In Brief Posted: 10:10 AM PST · February 13, 2026 Image Credits:Silas Stein/picture alliance / Getty Images Amanda Silberling OpenAI…
OpenAI sidesteps Nvidia with unusually fast coding model on plate-sized chips
Text settings Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more…
The Waymo World Model: A New Frontier for Autonomous Driving Simulation
February 6, 2026TechnologyChiyu Max Jiang, Xander Masotto, Bo SunShare on TwitterShare on LinkedInShare on FacebookThe Waymo Driver has traveled nearly…
Enterprises are measuring the wrong part of RAG
FeaturedVarun Raj February 1, 2026 CleoP made with MidjourneyEnterprises have moved quickly to adopt RAG to ground LLMs in proprietary…
Arcee’s U.S.-made, open source Trinity Large and 10T-checkpoint offer rare look at raw model intelligence
Carl Franzen January 30, 2026 Credit: VentureBeat made with Flux-1 on fal.aiSan Francisco-based AI lab Arcee made waves last year…
Moonshot AI debuts Kimi K2.5, most powerful open source LLM beating Opus 4.5 with swarm of parallel agents
FeaturedEmilia David January 27, 2026 Credit: VentureBeat made with MidjourneyChinese company Moonshot AI upgraded its open-sourced Kimi K2 model, transforming…
Conversational AI doesn’t understand users — ‘Intent First’ architecture does
FeaturedSreenivasa Reddy Hulebeedu Reddy January 25, 2026 Midjourney/VentureBeatThe modern customer has just one need that matters: Getting the thing they…
Show HN: Sweep, Open-weights 1.5B model for next-edit autocomplete
Sweep Next-Edit 1.5B (GGUF) A 1.5B parameter model for next-edit autocomplete, quantized to Q8_0 GGUF format. Model Description Sweep Next-Edit…
Why reinforcement learning plateaus without representation depth (and other key takeaways from NeurIPS 2025)
FeaturedMaitreyi Chatterjee,Devansh Agarwal January 17, 2026 CleoP made with MidjourneyImage generated using OpenAI’s DALL·E Every year, NeurIPS produces hundreds of…