AI Models News
Latest AI Models news and analysis — 95 articles tagged AI Models on The Robotics Media.
- Anthropic Says Claude Cracked A Nine-Loop Yang-Mills Amplitude, Beating A 2023 Human Record — — Anthropic said on 25 September that Claude Fable 5.1, running on ~96 CPUs for a week and $100-$2,000 of compute, computed the six-particle scattering...
- Google Ships Gemini 3.8 Flash TTS, 2,000+ Voices And SynthID Watermarks — — Google's Gemini 3.8 Flash TTS ships 2,000+ voices, promptable voice design and SynthID watermarks; Flash-Lite drops price another 50%.
- NVIDIA Open-Sources Nemotron 3 Diarization, Sets New VoiceArena Speaker-ID Record — — NVIDIA's ~100M-parameter open-weight diarization model beats the previous state of the art by 24% on VoiceArena.
- Meta Bets $1,300 On A VR Headset And Six New AI-Glasses SKUs — — At Meta Connect 2026, Zuckerberg unveiled a $1,300 tethered VR headset, a $449 Ray-Ban Meta Gen 3, $349 audio-only glasses and a Muse Charm keychain a...
- OpenAI Cuts GPT-6 Prices In Half With Sol And Luna Models — — OpenAI launches GPT-6 Sol and Luna at half the price of GPT-5.6, targeting professional coding, agentic work and high-volume clerical tasks.
- Perplexity Moves Comet's Browser-Control Feature Behind A Credits Paywall — — Comet's autonomous browser-driving mode is now billed in Computer credits, with Pro users getting 4,000 and Max users 35,000 plus a 10,000-a-month all...
- Figure AI's Helix 2.5 Cleans 30 Strangers' Homes It Has Never Seen — — A single frozen Helix 2.5 checkpoint hit 56% on tidying, towel-folding and bed-making across 30 unseen Bay Area homes, versus 9% without Index pretrai...
- Cohere and Aleph Alpha Sign Definitive Merger Agreement at ~$20B Valuation — — Cohere and Aleph Alpha have signed a definitive combination agreement, formalizing an April deal and creating a roughly $20B transatlantic sovereign-A...
- NVIDIA opens CUDA-Q Logical for fault-tolerant quantum — — NVIDIA has extended its CUDA-Q platform with an open-source layer for fault-tolerant quantum computing, adopted at launch by Fermilab, IQM, Infleqtion...
- Salesforce And NVIDIA Launch Koa, A CRM Reasoning Model Built On Nemotron And 27 Years Of Agentforce Data — — Salesforce and NVIDIA unveiled Koa at Dreamforce 2026 — a CRM-specific reasoning model post-trained from Nemotron 3 Super on synthetic scenarios spann...
- QuantrolOx Anchors UC Berkeley's Quantum Nexus With Five-Year MOU — — Oxford spin-out QuantrolOx becomes the first industry partner in UC Berkeley's Roger Herst Quantum Nexus, signing a five-year MOU to plug its Quantum...
- Inspiren Bags $70M Series C To Bring Physical AI To Senior Living — — NewView Capital leads a $70M Series C for Inspiren at a valuation above $500M as the New York startup scales its computer-vision fall detection stack...
- Skild AI Hits $100M Revenue Run Rate 10 Months After First Deploy — — Pittsburgh's Skild AI reveals a $100M annual revenue run rate and unveils S1, a video-taught robot foundation model built on NVIDIA infrastructure, te...
- IBM And NASA Ship Open-Source Lunar Foundation Model For Ice, Craters And Landing Sites — — NASA and IBM open-source the Lunar Foundation Model, joining the Prithvi family on Hugging Face and outperforming baselines by up to 23% on ice, crate...
- Positron Raises $875M Series C At $5B To Chase Nvidia On Inference — — NEA leads an $875M Series C for Positron at $5B, with Atreides, Valor Equity, Andra, SemiAnalysis Capital and Jim Clark backing the memory-first Asimo...
- Huawei Debuts Kirin 9050 Pro And Mate XT 2 Trifold, Days Before Apple's Foldable — — Huawei's Mate XT 2 trifold ships September 12 at RMB 19,999 with the Kirin 9050 Pro — the first commercial silicon built on LogicFolding 3D-stacked pa...
- DeepSeek Ships V4.1-Flash With 1M Context And 8x Smaller KV Cache — — DeepSeek releases V4.1-Flash under MIT with a 552B MoE backbone, 8B active parameters, FP4 KV cache and $0.15/M pricing that edges Opus 5 and GPT-5.6...
- HUMAIN Unveils humain-m3 Arabic Frontier Model Built On MiniMax M3 At LEAP 2026 — — PIF-backed HUMAIN launched humain-m3, a 428B Mixture-of-Experts Arabic language model post-trained on top of China's MiniMax M3, at LEAP 2026 in Riyad...
- Alibaba Ships Qwen3.8-Flash With 6B Active Params, Undercuts Rivals By 12x — — The 125B-parameter mixture-of-experts model activates just 6B per token, previews Qwen4 architecture and lists at $0.16 / $0.47 per million tokens.
- Higgsfield Bags $400M Series B At $5.4B As AI Video Revenue Hits $700M — — DST Global, Goldman Sachs and Intel back a $400 million round that quadruples Higgsfield's valuation to $5.4 billion as its annualized revenue jumps f...