Gemma news
The latest Gemma news, : Salesforce adds Koa Decisions with image and audio support to workflows.
Gemma is a lightweight AI model architecture developed by Google, referenced in several recent stories as the basis for on‑device and satellite AI applications. The coverage highlights a joint Google DeepMind‑GovTech on‑device scam‑detection model built on Gemma, Satlyt’s $8 million seed round to run Gemma on satellites, and related mentions of Gemma‑3‑4B in research datasets.
15 stories mentioning Gemma since 1 October 2026, newest first, each with its sources and discussion. Follow to see new ones on your front page.
-
Salesforce adds Koa Decisions with image and audio support to workflows
Salesforce launched Koa Decisions on November 6, 2026, a decision model built on Google’s open-weight Gemma that processes text, images, and audio. It replaces HyperClassifier, the predecessor embedded in Agentforce…
1 sourceUnite.AI -
Google invests $15-billion in Vizag AI hub
Google is investing $15-billion in a gigawatt‑scale AI hub in Vizag, India. The project includes undersea cable links, 125MW of solar power with ReNew, and 125MW of wind and solar power with CleanMax. Google plans to…
1 sourcebusinesstoday.in -
GitHub - W61k3r/LexiSuite-Releases: does qwen and gemma ok.
The project claims its engine, LexiEngine, matches or exceeds the speed of llama.cpp on a Radeon RX 7900 XTX GPU, with up to 30% faster prompt processing in some cases. Benchmarks from November 3–4, 2026, show…
1 source primary sourcegithub.com -
MIT researchers report GPT-6 Astra can read and write neural network weights, improving Qwen3.5-0.8B math score from 7.9% to 48.1%
MIT CSAIL researchers claim that GPT-6 Astra can natively read and write neural network weights without gradient descent or training loops, a capability they call Weight-Space Programming. In their experiments, Astra…
8 sources HN 41Latent SpaceThe Decodersimonwillison.netThe Rundown AI +3 more -
Opera adds WebAI standard for local browser AI with Google’s Gemma and Apple’s model
Opera introduced WebAI, a new web standard that lets users run AI models directly in their browser without cloud dependency. The feature supports open-weight models like Google’s Gemma and Apple’s on-device foundation…
2 sourcesUnite.AITechCrunch AI -
Ai2 releases FairStress benchmark to evaluate demographic fairness in LLM decisions
Ai2 introduced FairStress, a benchmark that measures hidden demographic fairness asymmetries in large language models. It contains 300 realistic two‑candidate decision scenarios across 29 domains such as hiring,…
1 source primary sourceHugging Face -
Opinion: AI scales human bias instead of eliminating it
AI systems inherit the judgments and blind spots embedded in the data they are trained on, so when they are used for hiring, healthcare or employee evaluation they can amplify existing discrimination at a speed no…
1 sourceUnite.AI -
Google DeepMind, Singapore unveil fraud detection AI; NVIDIA invests in four startups
Google DeepMind announced a partnership with Singapore’s GovTech to build an on‑device AI that flags social‑engineering scams in real time. The system runs entirely on the user’s device, using a fine‑tuned version of…
1 sourcenote.com -
Google DeepMind and Singapore's GovTech develop on-device scam-detecting AI
Google DeepMind and Singapore's GovTech announced on October 14, 2026 that they have jointly built an on-device artificial intelligence system that flags scam conversations without sending any audio to the cloud. The…
1 sourcenote.com -
New study finds AI struggles to correct false claims clearly across languages
A collaborative study by researchers from the University of Washington, Allen Institute for AI, LMU Munich, UC Berkeley, Stanford, and the University of Wisconsin-Madison examined how AI models handle factual errors in…
2 sourcesThe DecoderUnite.AI -
Mosaic learns sparse image-text latents for VLMs
Researchers introduced MOSAIC, a new method to understand vision-language models (VLMs). It creates a shared sparse latent space between images and text concepts. This enables localized influence attribution across the…
1 source primary sourcearXiv cs.AI -
Google DeepMind and GovTech announce on-device scam detection model for Singapore
Google DeepMind and Singapore's Government Technology Agency (GovTech) have unveiled an on-device scam detection model built on the lightweight Gemma architecture. The model runs locally on smartphones to identify…
4 sources primary sourcenote.comUnite.AIGoogle DeepMindchannelnewsasia.com -
Satlyt raises $8M to run AI on satellites with Google DeepMind’s Gemma
Satlyt, founded by former Google and SpaceX engineer Rama Afullo, has secured an $8 million seed round to develop software enabling AI workloads on satellites. The startup aims to create an open ecosystem for running…
1 source HN 5TechCrunch AI -
SymCE corpus released for Counterexample generation
SymCE is a new dataset of 4,707 false undergraduate‑algebra and real‑analysis conjectures, each paired with an executable Python verifier. The authors train Qwen3‑4B and Gemma‑3‑4B, showing that counterexample‑only…
1 source primary sourcearXiv cs.CL -
Satlyt raises $8M to run AI on satellites with Google’s Gemma model
Satlyt, founded by former Google and SpaceX engineer Rama Afullo, has secured an $8 million seed round to develop software that enables AI workloads on satellites. The company’s approach avoids building its own…
1 sourceTechCrunch AI
Questions about Gemma
What is the latest Gemma news?
As of 6 November 2026: Salesforce adds Koa Decisions with image and audio support to workflows. Salesforce launched Koa Decisions on November 6, 2026, a decision model built on Google’s open-weight Gemma that processes text, images, and audio.
What is Gemma?
Gemma is a lightweight AI model architecture developed by Google, referenced in several recent stories as the basis for on‑device and satellite AI applications. The coverage highlights a joint Google DeepMind‑GovTech on‑device scam‑detection model built on Gemma, Satlyt’s $8 million seed round to run Gemma on satellites, and related mentions of Gemma‑3‑4B in research datasets.