AI News — The Latent DailyThe latest AI news, every day: releases, launches, funding, policy and research.https://thelatentdaily.com/en-usOpenAI opens sponsored GPT-Rosalind access to eligible academic research labshttps://thelatentdaily.com/news/openai-rosalind-academic-plan/https://thelatentdaily.com/news/openai-rosalind-academic-plan/OpenAI introduced the Rosalind Academic Plan, offering eligible academic labs sponsored access to GPT-Rosalind and related tools for life-sciences research.Sun, 18 Oct 2026 08:52:00 GMTAI for Scienceopenaigpt-rosalindlife-sciencesacademiaresearchdario_admin@resolvia-crm.techLiquid AI releases LFM2.5-24B-A2B, a reasoning model for on-device agentshttps://thelatentdaily.com/news/liquid-ai-lfm25-24b-a2b-agents/https://thelatentdaily.com/news/liquid-ai-lfm25-24b-a2b-agents/Liquid AI introduced LFM2.5-24B-A2B, a 24B-parameter reasoning model with just 2B active parameters, built to run fast, private agents on local hardware.Sun, 18 Oct 2026 08:50:00 GMTModel Releasesliquid-ailfm2-5local-aiagentssmall-modelsdario_admin@resolvia-crm.techIBM ships Granite Switch 4.1, bundling chat, decision and speech models in one endpointhttps://thelatentdaily.com/news/ibm-granite-switch-4-1-release/https://thelatentdaily.com/news/ibm-granite-switch-4-1-release/IBM released Granite Switch 4.1 in 8B and 30B sizes, combining a base model with Decision and Guardian adapters and a speech recognition model in one deployment.Sun, 18 Oct 2026 08:47:00 GMTProduct Launchesibmgraniteenterprise-aispeechopen-weightsdario_admin@resolvia-crm.techNVIDIA RTX Spark laptops go on sale as Nemotron 3.5 Bolt 74B targets local agentshttps://thelatentdaily.com/news/nvidia-rtx-spark-laptops-launch/https://thelatentdaily.com/news/nvidia-rtx-spark-laptops-launch/NVIDIA said RTX Spark laptops became available October 16 with up to 128GB unified memory, alongside Nemotron 3.5 Bolt 74B, a model built for local AI agents.Sun, 18 Oct 2026 08:44:00 GMTAI PCs, Phones & Devicesnvidiartx-sparklocal-aiai-pcsnemotrondario_admin@resolvia-crm.techDeepMind and Singapore build an on-device Gemma model to flag conversational scamshttps://thelatentdaily.com/news/deepmind-singapore-on-device-scam-detection/https://thelatentdaily.com/news/deepmind-singapore-on-device-scam-detection/Google DeepMind and GovTech Singapore fine-tuned a lightweight Gemma model that analyzes chats on the phone itself to spot scams without sending data to the cloud.Sun, 18 Oct 2026 08:41:00 GMTPrivacy & Securitygoogle-deepmindgemmascamson-devicesingaporedario_admin@resolvia-crm.techDeepMind shows Antigravity agent workflow that surfaced an antimicrobial peptide leadhttps://thelatentdaily.com/news/deepmind-antigravity-antimicrobial-peptides/https://thelatentdaily.com/news/deepmind-antigravity-antimicrobial-peptides/Google DeepMind describes how biologists used its Antigravity platform to find a candidate antimicrobial peptide; the lead is now in wet-lab validation.Sun, 18 Oct 2026 08:38:00 GMTAI for Sciencegoogle-deepmindantigravitydrug-discoveryantibioticsbiologydario_admin@resolvia-crm.techBlack Forest Labs releases FLUX.3 Image Klein with bounding-box layout controlhttps://thelatentdaily.com/news/flux-3-image-klein-open-checkpoints/https://thelatentdaily.com/news/flux-3-image-klein-open-checkpoints/FLUX.3 Image Klein, an open image model from Black Forest Labs, handles text-to-image and multi-reference editing and adds layout control via bounding boxes.Sun, 18 Oct 2026 08:35:00 GMTModel Releasesblack-forest-labsflux-3image-generationdiffusersopen-weightsdario_admin@resolvia-crm.techMistral launches Timestral, its first family of time-series forecasting modelshttps://thelatentdaily.com/news/mistral-timestral-forecasting-models/https://thelatentdaily.com/news/mistral-timestral-forecasting-models/Mistral released Timestral, four zero-shot forecasting models from 5M to 209M parameters that it says rank first on the GIFT-Eval, TIME and fev-bench leaderboards.Sun, 18 Oct 2026 08:30:00 GMTModel Releasesmistral-aitimestralforecastingtime-seriesfoundation-modelsdario_admin@resolvia-crm.techMeta releases Muse Spark Open, an 816B open-weight multimodal model for codinghttps://thelatentdaily.com/news/meta-muse-spark-open-weights/https://thelatentdaily.com/news/meta-muse-spark-open-weights/Meta published Muse Spark Open, a mixture-of-experts model with roughly 816B parameters, 42B active per token, multimodal input and a 1M-token context window.Sun, 18 Oct 2026 08:00:00 GMTModel Releasesmetamuse-spark-openopen-weightsmultimodalcodingdario_admin@resolvia-crm.techReflection releases Beam, a 501B open-weight coding model under Apache 2.0https://thelatentdaily.com/news/reflection-beam-open-weight-model/https://thelatentdaily.com/news/reflection-beam-open-weight-model/Reflection, a new US lab, put out Beam: a 501B mixture-of-experts model with 23B active parameters, a 1M-token context and an Apache 2.0 license.Sun, 18 Oct 2026 07:30:00 GMTModel Releasesreflection-aibeamopen-weightscodingagentsdario_admin@resolvia-crm.tech