{"version":1,"generatedAt":"2026-09-18T17:26:32.131221Z","docs":"https://digestai.news/api","license":"Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse","category":{"slug":"models","name":"Generative AI & Models","url":"https://digestai.news/category/models"},"stories":[{"slug":"moonshot-ai-s-kimi-k3-opens-on-amazon-bedrock-with-2-8-trillion-parame","headline":"Moonshot AI's Kimi K3 opens on Amazon Bedrock with 2.8 trillion parameters","summary":"Moonshot AI announced that its Kimi K3 model is now available on Amazon Bedrock.","category":"models","firstPublishedAt":"2026-09-18T16:52:01Z","updatedAt":"2026-09-18T17:07:01Z","sourceCount":2,"hasPrimarySource":true,"url":"https://digestai.news/story/moonshot-ai-s-kimi-k3-opens-on-amazon-bedrock-with-2-8-trillion-parame","json":"https://digestai.news/story/moonshot-ai-s-kimi-k3-opens-on-amazon-bedrock-with-2-8-trillion-parame.json"},{"slug":"speculative-decoding-explained-how-ai-speeds-up-text-generation","headline":"Speculative decoding explained: how AI speeds up text generation","summary":"Speculative decoding speeds up autoregressive generation by letting a lightweight draft model propose a block of tokens that a larger target model verifies in parallel, without altering the target distribution.","category":"models","firstPublishedAt":"2026-09-18T12:00:00Z","updatedAt":"2026-09-18T12:00:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/speculative-decoding-explained-how-ai-speeds-up-text-generation","json":"https://digestai.news/story/speculative-decoding-explained-how-ai-speeds-up-text-generation.json"},{"slug":"alibaba-releases-qwen3-8-omni-flash-a-1m-token-omni-modal-model-with-a","headline":"Alibaba releases Qwen3.8-Omni-Flash, a 1M-token omni-modal model with agentic audio‑video","summary":"Alibaba’s Qwen team announced the launch of Qwen3.8-Omni-Flash, its first omni‑modal model that handles text, images, audio and video and returns text.","category":"models","firstPublishedAt":"2026-09-18T08:40:37Z","updatedAt":"2026-09-18T08:40:37Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/alibaba-releases-qwen3-8-omni-flash-a-1m-token-omni-modal-model-with-a","json":"https://digestai.news/story/alibaba-releases-qwen3-8-omni-flash-a-1m-token-omni-modal-model-with-a.json"},{"slug":"modality-discrepancy-transformer-reaches-0-7408-macro-f1-on-bah-datase","headline":"Modality Discrepancy Transformer reaches 0.7408 Macro F1 on BAH dataset","summary":"A new paper on arXiv proposes the Modality Discrepancy Transformer (MDT) to detect ambivalence and hesitancy in clinical videos, where facial, vocal, and linguistic cues conflict.","category":"models","firstPublishedAt":"2026-09-18T04:00:00Z","updatedAt":"2026-09-18T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/modality-discrepancy-transformer-reaches-0-7408-macro-f1-on-bah-datase","json":"https://digestai.news/story/modality-discrepancy-transformer-reaches-0-7408-macro-f1-on-bah-datase.json"},{"slug":"prismml-hopes-its-tiny-llm-will-change-how-we-all-use-ai","headline":"PrismML hopes its tiny LLM will change how we all use AI","summary":"PrismML hopes its tiny LLM will change how we all use AI If AI lab PrismML isn't on your radar yet, it should be.","category":"models","firstPublishedAt":"2026-09-17T22:34:09Z","updatedAt":"2026-09-17T22:34:09Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/prismml-hopes-its-tiny-llm-will-change-how-we-all-use-ai","json":"https://digestai.news/story/prismml-hopes-its-tiny-llm-will-change-how-we-all-use-ai.json"},{"slug":"openai-launches-astra-for-law-a-gpt-6-astra-variant-with-legal-search","headline":"OpenAI launches Astra for Law, a GPT-6 configuration for legal research","summary":"OpenAI announced Astra for Law on September 17, 2026.","category":"models","firstPublishedAt":"2026-09-17T21:04:02Z","updatedAt":"2026-09-18T13:42:43Z","sourceCount":7,"hasPrimarySource":false,"url":"https://digestai.news/story/openai-launches-astra-for-law-a-gpt-6-astra-variant-with-legal-search","json":"https://digestai.news/story/openai-launches-astra-for-law-a-gpt-6-astra-variant-with-legal-search.json"},{"slug":"freedomintelligence-releases-huatuogpt-3-9b-medical-llm-with-open-usage-guides","headline":"FreedomIntelligence releases HuatuoGPT-3-9B medical LLM with open usage guides","summary":"FreedomIntelligence has made its new HuatuoGPT-3-9B model publicly available on Hugging Face.","category":"models","firstPublishedAt":"2026-09-17T16:30:07Z","updatedAt":"2026-09-17T16:30:07Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/freedomintelligence-releases-huatuogpt-3-9b-medical-llm-with-open-usage-guides","json":"https://digestai.news/story/freedomintelligence-releases-huatuogpt-3-9b-medical-llm-with-open-usage-guides.json"},{"slug":"anthropic-says-ai-systems-lead-26-of-its-r-d","headline":"Anthropic says Claude leads 26% of its AI development work","summary":"Anthropic announced that its Claude model is now leading roughly a quarter of the company's internal AI research and development efforts.","category":"models","firstPublishedAt":"2026-09-17T13:36:00Z","updatedAt":"2026-09-18T07:30:00Z","sourceCount":6,"hasPrimarySource":false,"url":"https://digestai.news/story/anthropic-says-ai-systems-lead-26-of-its-r-d","json":"https://digestai.news/story/anthropic-says-ai-systems-lead-26-of-its-r-d.json"},{"slug":"openrouter-token-usage-soars-25-000-since-jan-2025","headline":"OpenRouter Token Usage Soars 25,000% Since Jan 2025","summary":"OpenRouter, a platform that lets developers plug AI models into their products, has seen an explosive rise in token consumption.","category":"models","firstPublishedAt":"2026-09-17T10:24:32Z","updatedAt":"2026-09-17T10:24:32Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/openrouter-token-usage-soars-25-000-since-jan-2025","json":"https://digestai.news/story/openrouter-token-usage-soars-25-000-since-jan-2025.json"},{"slug":"z-ai-deploys-glm5-3flash-on-100-000chip-chinese-cluster","headline":"Z.ai Deploys GLM‑5.3‑Flash on 100,000‑Chip Chinese Cluster","summary":"Z.ai announced a technical account detailing how it built a production‑grade inference service for its GLM‑5.3‑Flash model on a cluster of more than 100,000 Chinese‑made AI accelerators, a first at this scale.","category":"models","firstPublishedAt":"2026-09-17T09:56:01Z","updatedAt":"2026-09-17T09:56:01Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/z-ai-deploys-glm5-3flash-on-100-000chip-chinese-cluster","json":"https://digestai.news/story/z-ai-deploys-glm5-3flash-on-100-000chip-chinese-cluster.json"},{"slug":"google-launches-gemini-3-8-live-voice-model-with-0-023-per-minute-pric","headline":"Google launches Gemini 3.8 Live voice model with $0.023 per minute pricing","summary":"Google announced Gemini 3.8 Live on September 15, 2026, a voice AI model that enables continuous, natural conversation in 97 languages and can run background tasks such as search and tool invocation.","category":"models","firstPublishedAt":"2026-09-17T07:25:00Z","updatedAt":"2026-09-17T07:25:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/google-launches-gemini-3-8-live-voice-model-with-0-023-per-minute-pric","json":"https://digestai.news/story/google-launches-gemini-3-8-live-voice-model-with-0-023-per-minute-pric.json"},{"slug":"faircompressagent-an-agentic-framework-for-fairness-aware-model-compression","headline":"FairCompressAgent: An Agentic Framework for Fairness-Aware Model Compression","summary":"This paper introduces FairCompressAgent (FCA), a novel agentic framework designed to handle the complexities of model compression while maintaining fairness and minimizing resource usage.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/faircompressagent-an-agentic-framework-for-fairness-aware-model-compression","json":"https://digestai.news/story/faircompressagent-an-agentic-framework-for-fairness-aware-model-compression.json"},{"slug":"llms-outperform-traditional-chinese-medicine-physicians-in-case-evaluations","headline":"LLMs outperform traditional Chinese medicine physicians in case evaluations","summary":"A study involving 16 large language models (LLMs) and 60 practicing TCM physicians evaluated real-world cases from 62 hospitals.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/llms-outperform-traditional-chinese-medicine-physicians-in-case-evaluations","json":"https://digestai.news/story/llms-outperform-traditional-chinese-medicine-physicians-in-case-evaluations.json"},{"slug":"non-standard-english-queries-routinely-sent-to-lower-capacity-llms-study-finds","headline":"Non-Standard English Queries Routinely Sent to Lower-Capacity LLMs, Study Finds","summary":"Researchers examined how large‑language‑model (LLM) services route user queries to different model tiers based on a cheap estimate of query complexity.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/non-standard-english-queries-routinely-sent-to-lower-capacity-llms-study-finds","json":"https://digestai.news/story/non-standard-english-queries-routinely-sent-to-lower-capacity-llms-study-finds.json"},{"slug":"llm-response-distortion-across-dark-triad-traits","headline":"LLM Response Distortion Across Dark Triad Traits","summary":"A study examines how seven state-of-the-art Large Language Models (LLMs) modulate the expression of Machiavellianism, narcissism, and psychopathy under fake-good and fake-bad conditions.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/llm-response-distortion-across-dark-triad-traits","json":"https://digestai.news/story/llm-response-distortion-across-dark-triad-traits.json"},{"slug":"sage-streamlines-enterprise-document-conversion","headline":"SAGE: Streamlines Enterprise Document Conversion","summary":"Enterprise guideline documents often require manual effort to convert into structured work artifacts, a process that can take days.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/sage-streamlines-enterprise-document-conversion","json":"https://digestai.news/story/sage-streamlines-enterprise-document-conversion.json"},{"slug":"gvd-a-unified-framework-for-document-versioning","headline":"GVD: A Unified Framework for Document Versioning","summary":"Document repositories often contain multiple versions of the same content due to revisions and updates.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/gvd-a-unified-framework-for-document-versioning","json":"https://digestai.news/story/gvd-a-unified-framework-for-document-versioning.json"},{"slug":"graphecho-reveals-agent-overlooking-evidence","headline":"GraphEcho Reveals Agent Overlooking Evidence","summary":"A new study, 'GraphEcho,' explores how large language model (LLM) agents can traverse more paths in graphs without acquiring additional evidence.","category":"models","firstPublishedAt":"2026-09-17T04:00:00Z","updatedAt":"2026-09-17T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/graphecho-reveals-agent-overlooking-evidence","json":"https://digestai.news/story/graphecho-reveals-agent-overlooking-evidence.json"},{"slug":"anthropic-s-invisible-text-watermarking-in-claude-ai","headline":"Anthropic's Invisible Text Watermarking in Claude AI","summary":"Anthropic has introduced a watermarking system for their Claude AI text generation tool to identify AI-generated content more reliably.","category":"models","firstPublishedAt":"2026-09-17T03:47:00Z","updatedAt":"2026-09-17T18:52:02Z","sourceCount":4,"hasPrimarySource":false,"url":"https://digestai.news/story/anthropic-s-invisible-text-watermarking-in-claude-ai","json":"https://digestai.news/story/anthropic-s-invisible-text-watermarking-in-claude-ai.json"},{"slug":"ai-chatbots-lag-behind-on-recent-information","headline":"AI Chatbots Lag Behind on Recent Information","summary":"When testing popular AI chatbots like ChatGPT, Claude, and Gemini, I found that they often lack up-to-date knowledge about recent events.","category":"models","firstPublishedAt":"2026-09-16T23:15:00Z","updatedAt":"2026-09-16T23:15:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/ai-chatbots-lag-behind-on-recent-information","json":"https://digestai.news/story/ai-chatbots-lag-behind-on-recent-information.json"},{"slug":"gliformer-a-new-encoder-framework-hits-high-nested-json-extraction-f1-score","headline":"GLiFormer: A New Encoder Framework Hits High Nested JSON Extraction F1 Score","summary":"Knowledgator Engineering has released GLiFormer, a schema-conditioned encoder framework designed to handle various information extraction tasks such as named-entity recognition (NER), text classification, relation extraction, and nested…","category":"models","firstPublishedAt":"2026-09-16T21:19:32Z","updatedAt":"2026-09-16T21:19:32Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/gliformer-a-new-encoder-framework-hits-high-nested-json-extraction-f1-score","json":"https://digestai.news/story/gliformer-a-new-encoder-framework-hits-high-nested-json-extraction-f1-score.json"},{"slug":"aigenerated-film-odysseus-the-fall-disappoints-with-technical-flaws","headline":"AI‑generated film 'Odysseus: The Fall' disappoints with technical flaws","summary":"Fountain 0, an AI startup, released Odysseus: The Fall, a 2.5‑hour movie claimed to be the first fully AI‑generated blockbuster.","category":"models","firstPublishedAt":"2026-09-16T20:59:13Z","updatedAt":"2026-09-16T20:59:13Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/aigenerated-film-odysseus-the-fall-disappoints-with-technical-flaws","json":"https://digestai.news/story/aigenerated-film-odysseus-the-fall-disappoints-with-technical-flaws.json"},{"slug":"google-moves-chrome-to-a-2week-update-cycle-to-narrow-patch-gap","headline":"Google moves Chrome to a 2‑week update cycle to narrow patch gap","summary":"Google has shortened the release cadence for its Chrome browser from four weeks to two weeks, a change announced in a blog post and confirmed by Chrome lead Ben Mason.","category":"models","firstPublishedAt":"2026-09-16T16:07:00Z","updatedAt":"2026-09-16T16:07:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/google-moves-chrome-to-a-2week-update-cycle-to-narrow-patch-gap","json":"https://digestai.news/story/google-moves-chrome-to-a-2week-update-cycle-to-narrow-patch-gap.json"},{"slug":"llama-cpp-b11003-adds-hrm-text-support-for-dfm-mimir-1b-model","headline":"llama.cpp b11003 adds HRM-Text support for DFM Mimir 1B model","summary":"The llama.cpp project released version b11003, introducing support for the HRM‑Text architecture used by the DFM Mimir 1B model.","category":"models","firstPublishedAt":"2026-09-16T15:51:51Z","updatedAt":"2026-09-16T15:51:51Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/llama-cpp-b11003-adds-hrm-text-support-for-dfm-mimir-1b-model","json":"https://digestai.news/story/llama-cpp-b11003-adds-hrm-text-support-for-dfm-mimir-1b-model.json"},{"slug":"scikit-llm-cheat-sheet-integrating-language-models","headline":"Scikit-LLM Cheat Sheet: Integrating Language Models","summary":"This article introduces Scikit-LLM, which wraps language models in the scikit-learn estimator API.","category":"models","firstPublishedAt":"2026-09-16T13:58:19Z","updatedAt":"2026-09-16T13:58:19Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/scikit-llm-cheat-sheet-integrating-language-models","json":"https://digestai.news/story/scikit-llm-cheat-sheet-integrating-language-models.json"},{"slug":"model-serves-memory-kv-cache-tax-in-ai-inference","headline":"Model Serves Memory: KV Cache Tax in AI Inference","summary":"A mid-sized model deployment fell over due to a memory issue when under real traffic, despite having enough weights and compute resources.","category":"models","firstPublishedAt":"2026-09-16T12:30:02Z","updatedAt":"2026-09-16T12:30:02Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/model-serves-memory-kv-cache-tax-in-ai-inference","json":"https://digestai.news/story/model-serves-memory-kv-cache-tax-in-ai-inference.json"},{"slug":"brin-returns-to-google-s-ai-research","headline":"Brin Returns to Google's AI Research","summary":"Google co-founder Sergey Brin is once again taking a hands-on role in the company’s artificial intelligence efforts.","category":"models","firstPublishedAt":"2026-09-16T10:00:00Z","updatedAt":"2026-09-16T10:00:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/brin-returns-to-google-s-ai-research","json":"https://digestai.news/story/brin-returns-to-google-s-ai-research.json"},{"slug":"ggml-org-merges-hc-ops-into-qwen4exp-graph-in-llama-cpp","headline":"ggml-org Merges hc Ops into qwen4exp Graph in llama.cpp","summary":"A recent pull request (28901) in the open‑source llama.cpp repository has added new high‑capacity (hc) operations to the qwen4exp graph, targeting both CPU and CUDA backends.","category":"models","firstPublishedAt":"2026-09-16T09:51:41Z","updatedAt":"2026-09-16T09:51:41Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/ggml-org-merges-hc-ops-into-qwen4exp-graph-in-llama-cpp","json":"https://digestai.news/story/ggml-org-merges-hc-ops-into-qwen4exp-graph-in-llama-cpp.json"},{"slug":"one-extension-could-hijack-multiple-ai-assistants","headline":"One Extension Could Hijack Multiple AI Assistants","summary":"Security researchers at Forever Security discovered that one ordinary browser extension could control multiple AI assistants built into five Chromium-based products: Gemini Live in Chrome, Perplexity Comet, Microsoft Edge, Opera Neon, and…","category":"models","firstPublishedAt":"2026-09-16T07:36:00Z","updatedAt":"2026-09-16T07:36:00Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/one-extension-could-hijack-multiple-ai-assistants","json":"https://digestai.news/story/one-extension-could-hijack-multiple-ai-assistants.json"},{"slug":"prior-labs-releases-tabpfn-3-5-beating-2015-kaggle-winner-with-default-settings","headline":"Prior Labs releases TabPFN-3.5, beating 2015 Kaggle winner with default settings","summary":"Prior Labs has launched TabPFN-3.5, a 220M-parameter tabular foundation model that outperforms the winning solution of the 2015 Otto Group Kaggle competition.","category":"models","firstPublishedAt":"2026-09-16T06:52:18Z","updatedAt":"2026-09-16T06:52:18Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/prior-labs-releases-tabpfn-3-5-beating-2015-kaggle-winner-with-default-settings","json":"https://digestai.news/story/prior-labs-releases-tabpfn-3-5-beating-2015-kaggle-winner-with-default-settings.json"},{"slug":"nums-ai-launches-causilo-a-tabular-foundation-model-that-tops-tabarena","headline":"Nums AI launches Causilo, a tabular foundation model that tops TabArena benchmarks","summary":"Nums AI released Causilo, a pretrained tabular foundation model for classification and regression, with a scikit‑learn‑style API, Apache‑2.0 code, and weights hosted on Hugging Face.","category":"models","firstPublishedAt":"2026-09-16T04:59:47Z","updatedAt":"2026-09-16T04:59:47Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/nums-ai-launches-causilo-a-tabular-foundation-model-that-tops-tabarena","json":"https://digestai.news/story/nums-ai-launches-causilo-a-tabular-foundation-model-that-tops-tabarena.json"},{"slug":"mlabonne-releases-lfm2-5-230m-chess-a-230m-parameter-chess-engine-language-model","headline":"mlabonne releases LFM2.5-230M-Chess, a 230M-parameter chess engine language model","summary":"A new open‑weight model, LFM2.5-230M-Chess, has been added to Hugging Face by the mlabonne community.","category":"models","firstPublishedAt":"2026-09-16T04:53:29Z","updatedAt":"2026-09-16T04:53:29Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/mlabonne-releases-lfm2-5-230m-chess-a-230m-parameter-chess-engine-language-model","json":"https://digestai.news/story/mlabonne-releases-lfm2-5-230m-chess-a-230m-parameter-chess-engine-language-model.json"},{"slug":"gpu-cache-placement-insights-for-efficient-sessions","headline":"GPU Cache Placement: Insights for Efficient Sessions","summary":"A new study explores how to best allocate key-value (KV) caches across different memory tiers—Graphics Processing Unit (GPU), CPU, and Solid State Drive (SSD)—to optimize session management in AI systems.","category":"models","firstPublishedAt":"2026-09-16T04:00:00Z","updatedAt":"2026-09-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/gpu-cache-placement-insights-for-efficient-sessions","json":"https://digestai.news/story/gpu-cache-placement-insights-for-efficient-sessions.json"},{"slug":"safe-error-correction-for-language-models","headline":"Safe Error Correction for Language Models","summary":"Researchers have developed CRN v2, a lightweight logit-level correction module that can fix errors in frozen language models without degrading their base capabilities.","category":"models","firstPublishedAt":"2026-09-16T04:00:00Z","updatedAt":"2026-09-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/safe-error-correction-for-language-models","json":"https://digestai.news/story/safe-error-correction-for-language-models.json"},{"slug":"cadworld-a-new-benchmark-for-long-horizon-computer-aided-design","headline":"CADWorld: A New Benchmark for Long-Horizon Computer-Aided Design","summary":"Computer-use agents are increasingly evaluated in realistic desktop environments, but benchmarks often lack coverage of professional engineering workflows.","category":"models","firstPublishedAt":"2026-09-16T04:00:00Z","updatedAt":"2026-09-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/cadworld-a-new-benchmark-for-long-horizon-computer-aided-design","json":"https://digestai.news/story/cadworld-a-new-benchmark-for-long-horizon-computer-aided-design.json"},{"slug":"calibrated-router-boosts-llm-serving-efficiency","headline":"Calibrated Router Boosts LLM Serving Efficiency","summary":"A study examines a new approach to routing requests in disaggregated Large Language Model (LLM) serving systems.","category":"models","firstPublishedAt":"2026-09-16T04:00:00Z","updatedAt":"2026-09-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/calibrated-router-boosts-llm-serving-efficiency","json":"https://digestai.news/story/calibrated-router-boosts-llm-serving-efficiency.json"},{"slug":"llms-detect-pain-in-text","headline":"LLMs Detect Pain in Text","summary":"A recent study published on arXiv has found that large language models (LLMs) can detect pain in text, distinguishing it from other negative emotions like fear or sadness.","category":"models","firstPublishedAt":"2026-09-16T04:00:00Z","updatedAt":"2026-09-16T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/llms-detect-pain-in-text","json":"https://digestai.news/story/llms-detect-pain-in-text.json"},{"slug":"nvidia-cudnn-graph-api-tutorial-fusion-autotuning","headline":"NVIDIA cuDNN Graph API Tutorial: Fusion & Autotuning","summary":"This tutorial walks through NVIDIA’s cuDNN Frontend graph API.","category":"models","firstPublishedAt":"2026-09-15T21:37:11Z","updatedAt":"2026-09-15T21:37:11Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/nvidia-cudnn-graph-api-tutorial-fusion-autotuning","json":"https://digestai.news/story/nvidia-cudnn-graph-api-tutorial-fusion-autotuning.json"},{"slug":"zerodependency-1-8-mb-numberwang-neural-net-classifies-numbers-in-11-languages","headline":"Zero‑dependency 1.8 MB Numberwang neural net classifies numbers in 11 languages","summary":"A tiny neural network called Numberwang, released on GitHub, can decide whether a string represents a “Numberwang” or a “Wangernumb” in eleven languages.","category":"models","firstPublishedAt":"2026-09-15T19:27:35Z","updatedAt":"2026-09-15T19:27:35Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/zerodependency-1-8-mb-numberwang-neural-net-classifies-numbers-in-11-languages","json":"https://digestai.news/story/zerodependency-1-8-mb-numberwang-neural-net-classifies-numbers-in-11-languages.json"},{"slug":"typesafe-ai-unveils-jev-a-frontier-model-up-to-400-cheaper-and-200-faster","headline":"TypeSafe AI launches Jev, a decision model claimed up to 193.6x faster and 444.6x cheaper","summary":"TypeSafe AI announced the release of Jev, its first “System One” model that returns typed probabilistic decisions instead of free‑form text.","category":"models","firstPublishedAt":"2026-09-15T19:25:03Z","updatedAt":"2026-09-17T03:12:05Z","sourceCount":9,"hasPrimarySource":true,"url":"https://digestai.news/story/typesafe-ai-unveils-jev-a-frontier-model-up-to-400-cheaper-and-200-faster","json":"https://digestai.news/story/typesafe-ai-unveils-jev-a-frontier-model-up-to-400-cheaper-and-200-faster.json"},{"slug":"ai-actor-tilly-norwood-can-t-engage-in-current-events","headline":"AI Actor Tilly Norwood Can't Engage in Current Events","summary":"Particle 6 Group developed an AI-generated character named Tilly Norwood to promote its film projects.","category":"models","firstPublishedAt":"2026-09-15T18:00:09Z","updatedAt":"2026-09-15T18:00:09Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/ai-actor-tilly-norwood-can-t-engage-in-current-events","json":"https://digestai.news/story/ai-actor-tilly-norwood-can-t-engage-in-current-events.json"},{"slug":"google-deepmind-launches-gemini-3-8-live-and-3-8-live-extended-thinking-models","headline":"Google launches Gemini 3.8 Live and Extended Thinking voice models","summary":"Google announced on September 15, 2026 the release of two speech‑to‑speech models—Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking—designed for real‑time voice agents.","category":"models","firstPublishedAt":"2026-09-15T17:05:57Z","updatedAt":"2026-09-17T13:02:21Z","sourceCount":7,"hasPrimarySource":true,"url":"https://digestai.news/story/google-deepmind-launches-gemini-3-8-live-and-3-8-live-extended-thinking-models","json":"https://digestai.news/story/google-deepmind-launches-gemini-3-8-live-and-3-8-live-extended-thinking-models.json"},{"slug":"novo-s-michael-rangel-on-the-transformative-role-of-ai-in-fintech","headline":"Novo's Michael Rangel on the Transformative Role of AI in Fintech","summary":"Michael Rangel, founder of fintech platform Novo, discusses how artificial intelligence (AI) has revolutionized his company.","category":"models","firstPublishedAt":"2026-09-15T16:46:32Z","updatedAt":"2026-09-15T16:46:32Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/novo-s-michael-rangel-on-the-transformative-role-of-ai-in-fintech","json":"https://digestai.news/story/novo-s-michael-rangel-on-the-transformative-role-of-ai-in-fintech.json"},{"slug":"amazon-sagemaker-serverless-customization-fine-tunes-qwen3-8b-for-product","headline":"Amazon SageMaker serverless customization fine-tunes Qwen3-8B for product tagging","summary":"Retail catalogs often arrive with inconsistent attributes, making manual tagging of thousands of SKUs slow and error‑prone.","category":"models","firstPublishedAt":"2026-09-15T16:11:36Z","updatedAt":"2026-09-15T16:11:36Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/amazon-sagemaker-serverless-customization-fine-tunes-qwen3-8b-for-product","json":"https://digestai.news/story/amazon-sagemaker-serverless-customization-fine-tunes-qwen3-8b-for-product.json"},{"slug":"amazon-sagemaker-ai-adds-instance-preference-lists","headline":"Amazon SageMaker AI Adds Instance Preference Lists","summary":"AWS has introduced Instance preference lists for Amazon SageMaker AI Training Jobs and Processing Jobs, allowing users to specify a list of up to five instance types in order of preference.","category":"models","firstPublishedAt":"2026-09-15T16:01:47Z","updatedAt":"2026-09-15T16:01:47Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/amazon-sagemaker-ai-adds-instance-preference-lists","json":"https://digestai.news/story/amazon-sagemaker-ai-adds-instance-preference-lists.json"},{"slug":"google-expands-ai-language-tools-to-300-languages-with-new-gemini-models","headline":"Google expands AI language tools to 300+ languages with new Gemini models","summary":"Google has announced significant advancements in its multilingual AI capabilities, stating that its technologies now support over 300 languages and serve 7 billion people.","category":"models","firstPublishedAt":"2026-09-15T16:00:00Z","updatedAt":"2026-09-15T16:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/google-expands-ai-language-tools-to-300-languages-with-new-gemini-models","json":"https://digestai.news/story/google-expands-ai-language-tools-to-300-languages-with-new-gemini-models.json"},{"slug":"no-big-deal-a-boring-ai-generated-sitcom","headline":"No Big Deal: A Boring AI-Generated Sitcom","summary":"The article discusses the debut of an AI-generated sitcom called 'No Big Deal'.","category":"models","firstPublishedAt":"2026-09-15T15:00:53Z","updatedAt":"2026-09-15T15:00:53Z","sourceCount":1,"hasPrimarySource":false,"url":"https://digestai.news/story/no-big-deal-a-boring-ai-generated-sitcom","json":"https://digestai.news/story/no-big-deal-a-boring-ai-generated-sitcom.json"},{"slug":"salesforce-launches-koa-a-reasoning-model-built-on-nvidia-s-nemotron","headline":"Salesforce launches Koa, its first CRM reasoning model built on NVIDIA Nemotron 3 Super","summary":"Salesforce has officially unveiled Koa, its first proprietary CRM reasoning model, developed in partnership with NVIDIA.","category":"models","firstPublishedAt":"2026-09-15T12:00:00Z","updatedAt":"2026-09-16T23:13:45Z","sourceCount":6,"hasPrimarySource":true,"url":"https://digestai.news/story/salesforce-launches-koa-a-reasoning-model-built-on-nvidia-s-nemotron","json":"https://digestai.news/story/salesforce-launches-koa-a-reasoning-model-built-on-nvidia-s-nemotron.json"},{"slug":"zgcm-1-a-new-open-math-search-foundation-model","headline":"ZGCM-1: A New Open Math & Search Foundation Model","summary":"In this work, researchers have unveiled ZGCM-1, a groundbreaking 7B dense foundation model designed for mathematical reasoning and agentic search.","category":"models","firstPublishedAt":"2026-09-15T04:00:00Z","updatedAt":"2026-09-15T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/zgcm-1-a-new-open-math-search-foundation-model","json":"https://digestai.news/story/zgcm-1-a-new-open-math-search-foundation-model.json"},{"slug":"hybrid-ai-framework-improves-medical-text-summarization","headline":"Hybrid AI Framework Improves Medical Text Summarization","summary":"A new hybrid framework combining CNN and LSTM models has been developed to improve the extraction of relevant sentences from biomedical and clinical texts.","category":"models","firstPublishedAt":"2026-09-15T04:00:00Z","updatedAt":"2026-09-15T04:00:00Z","sourceCount":1,"hasPrimarySource":true,"url":"https://digestai.news/story/hybrid-ai-framework-improves-medical-text-summarization","json":"https://digestai.news/story/hybrid-ai-framework-improves-medical-text-summarization.json"}]}