{"apiVersion":"v1","taxonomyVersion":1,"generatedAt":"2026-10-29T12:39:14.448Z","data":{"id":"brf_rXhN8Ge2cs91","kind":"daily","slug":"2026-10-28","window":{"from":"2026-10-28T00:00:00.000Z","to":"2026-10-29T00:00:00.000Z"},"version":1,"quiet":false,"title":"AI brief · 28 October 2026","headline":{"text":"New models bring live speech translation and voice features","recordIds":["wsa_v989cwdcn7ub","wsa_uvvu8wqvjfxj","wsa_7uraa6h3vdhc","wsa_3rqvsdzbskn8"]},"dek":{"text":"New speech models can handle live multilingual conversation, including continuous Chinese–English translation and real-time speech apps with web search or function calling. Several new models focus on reading documents, from extracting receipt totals to parsing multi-page files and tables. Liquid AI released models for on-device voice and video tasks, as well as local agents and media classification. Maya 2 Native offers streaming speech generation in 12 languages, with open weights and voice cloning across languages.","recordIds":["wsa_v989cwdcn7ub","wsa_uvvu8wqvjfxj","wsa_7uraa6h3vdhc","wsa_rv5dhcpijwn6","wsa_e4xixqvv53t5","wsa_26mmc26i8b8p","wsa_guk7j9k3f7bu","wsa_ug7p3suzjfva","wsa_3rqvsdzbskn8"]},"tldr":[{"text":"New speech models can handle live multilingual conversation, including continuous Chinese–English translation and real-time speech apps with web search or function calling.","recordIds":["wsa_v989cwdcn7ub","wsa_uvvu8wqvjfxj","wsa_7uraa6h3vdhc"]},{"text":"Several new models focus on reading documents, from extracting receipt totals to parsing multi-page files and tables.","recordIds":["wsa_rv5dhcpijwn6","wsa_e4xixqvv53t5"]},{"text":"Liquid AI released models for on-device voice and video tasks, as well as local agents and media classification.","recordIds":["wsa_26mmc26i8b8p","wsa_guk7j9k3f7bu","wsa_ug7p3suzjfva"]},{"text":"Maya 2 Native offers streaming speech generation in 12 languages, with open weights and voice cloning across languages.","recordIds":["wsa_3rqvsdzbskn8"]}],"stories":[{"blocks":[{"kind":"paragraph","items":[{"text":"**Hy-SimulMT-30B-A3B** and **Hy-SimulMT-30B-A3B-FP8** translate incoming Chinese or English speech into text and spoken translation as a conversation unfolds. Their voice-cloning feature can preserve the speaker’s vocal characteristics, but it requires a configured text-to-speech service.","recordIds":["wsa_v989cwdcn7ub","wsa_uvvu8wqvjfxj"]},{"text":"**Qwen-Audio-3.1-Realtime-Flash** is built for two-way speech interactions: it can take in and produce both text and audio, and supports web search and function calling. Developers could use those tools to build speech apps that look up information or trigger actions during a conversation.","recordIds":["wsa_7uraa6h3vdhc"]}]},{"kind":"paragraph","items":[{"text":"**Maya 2 Native** generates streaming speech in 12 languages and includes 20 voices. Its open weights and inference code can support locally hosted voice applications, while cross-lingual voice cloning lets it generate speech in another language using a voice sample.","recordIds":["wsa_3rqvsdzbskn8"]}]}],"heading":"New models bring live speech translation and voice features"},{"blocks":[{"kind":"paragraph","items":[{"text":"**MiniCPM-OCR 1.0** is designed to read more than plain text: it can handle tables, formulas, charts and other document elements in Chinese and English. One practical use is extracting a specific field, such as the total on a receipt.","recordIds":["wsa_rv5dhcpijwn6"]},{"text":"**Unlimited OCR 2** is aimed at parsing images and long, multi-page documents, including PDFs. Its description says it can process documents beyond a finite context window, but does not give a specific page limit.","recordIds":["wsa_e4xixqvv53t5"]}]}],"heading":"New OCR models target receipts, tables and long documents"},{"blocks":[{"kind":"paragraph","items":[{"text":"**LFM3-Omni-2B** can take in text, images, video and audio, then return text, speech or both. The model is designed for on-device use; listed applications include transcribing or translating speech and asking questions about a video’s visuals or soundtrack.","recordIds":["wsa_26mmc26i8b8p"]},{"text":"**d1-Omni-2B** is a decision model that can classify or extract information from text and media, such as locating objects in images or routing calls based on audio. Instead of generating a written response, it returns answers with probabilities.","recordIds":["wsa_guk7j9k3f7bu"]},{"text":"**LFM3-1.7B** is an open-weight text model intended for local agents and specialized applications. **LFM3-1.7B-Base** is a separate pre-trained checkpoint intended as a starting point for substantial fine-tuning.","recordIds":["wsa_ug7p3suzjfva","wsa_9yeug8as48fe"]}]}],"heading":"Liquid AI adds models for local media tasks and agents"}],"radar":[{"text":"**Ling-3.1-flash-base** is an open-weight base model for continued pretraining, adapting a model to a domain, and other research. It is intended as a starting checkpoint rather than a finished assistant.","recordIds":["wsa_mj447cwhqjya"]},{"text":"**Chaos Intelligence** is a programming language and development environment for building programs that combine functions, language models and agent workflows.","recordIds":["wsa_dpenhp6w8mt5"]}],"sections":[{"heading":"Models","category":"model","recordIds":["wsa_mj447cwhqjya","wsa_3rqvsdzbskn8","wsa_7uraa6h3vdhc","wsa_26mmc26i8b8p","wsa_guk7j9k3f7bu","wsa_ug7p3suzjfva","wsa_9yeug8as48fe","wsa_rv5dhcpijwn6","wsa_v989cwdcn7ub","wsa_uvvu8wqvjfxj","wsa_e4xixqvv53t5"],"sentences":[]},{"heading":"Developer platforms","category":"developer platform","recordIds":["wsa_dpenhp6w8mt5"],"sentences":[]}],"items":[{"recordId":"wsa_mj447cwhqjya","versionId":"ver_9S2rJ3hrDzYA","eventType":"new_development","category":"model","title":"Ling-3.1-flash-base","maker":"inclusionAI","summary":"Ling-3.1-flash-base is a base checkpoint released to support research and community-driven innovation. It has 560B total parameters and 25B active parameters per token, with a hybrid-linear Mixture-of-Experts architecture.","releaseNote":null,"classifications":{"developmentType":"model","capability":["text generation","code generation","training","fine-tuning"],"modality":["text","code"],"access":["open weights"],"deployment":["on-premises"],"pricing":["unknown"],"targetUser":["researcher","developer"]},"citations":[{"url":"https://huggingface.co/inclusionAI/Ling-3.1-flash-base","authority":"owner-controlled","publisher":"Hugging Face"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-25T02:15:27.000Z","effectiveAt":"2026-10-28T00:21:06.064Z","url":"https://wowsuchai.withglint.app/records/wsa_mj447cwhqjya","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_mj447cwhqjya","webUrl":"https://wowsuchai.withglint.app/records/wsa_mj447cwhqjya"},{"recordId":"wsa_3rqvsdzbskn8","versionId":"ver_Wf0e7BNVamIa","eventType":"new_development","category":"model","title":"Maya 2 Native","maker":"Maya Research","summary":"Maya 2 Native is a streaming text-to-speech model that generates 24 kHz speech in 12 languages, with 20 built-in voices. It includes open weights and inference code, supports zero-shot cross-lingual voice cloning, and is released under Apache-2.0.","releaseNote":null,"classifications":{"developmentType":"model","capability":["audio generation"],"modality":["text","audio"],"access":["open weights","API","SDK"],"deployment":["on-premises"],"pricing":["open-source"],"targetUser":["developer"]},"citations":[{"url":"https://huggingface.co/maya-research/maya2-native","authority":"owner-controlled","publisher":"Hugging Face"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-26T02:58:10.000Z","effectiveAt":"2026-10-28T00:21:01.932Z","url":"https://wowsuchai.withglint.app/records/wsa_3rqvsdzbskn8","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_3rqvsdzbskn8","webUrl":"https://wowsuchai.withglint.app/records/wsa_3rqvsdzbskn8"},{"recordId":"wsa_7uraa6h3vdhc","versionId":"ver_WXAjLTXEKLpZ","eventType":"new_development","category":"model","title":"Qwen-Audio-3.1-Realtime-Flash","maker":"Alibaba Qwen","summary":"Qwen-Audio-3.1-Realtime-Flash is a real-time duplex speech model that accepts and produces text and audio. It supports function calling, web search, voice cloning, and a 262,144-token context window.","releaseNote":null,"classifications":{"developmentType":"model","capability":["text generation","audio generation","search","agents"],"modality":["text","audio"],"access":["API"],"deployment":["hosted"],"pricing":["unknown"],"targetUser":["developer"]},"citations":[{"url":"https://www.alibabacloud.com/help/en/model-studio/newly-released-models#2026-10-27","authority":"owner-controlled","publisher":"Alibaba Qwen"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-27T00:00:00.000Z","effectiveAt":"2026-10-28T00:20:55.994Z","url":"https://wowsuchai.withglint.app/records/wsa_7uraa6h3vdhc","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_7uraa6h3vdhc","webUrl":"https://wowsuchai.withglint.app/records/wsa_7uraa6h3vdhc"},{"recordId":"wsa_26mmc26i8b8p","versionId":"ver_BjrnskvwLjPl","eventType":"new_development","category":"model","title":"LFM3-Omni-2B","maker":"Liquid AI","summary":"LFM3-Omni-2B is a post-trained omni-modal model that accepts text, images, video, and audio, and produces text, speech, or interleaved text and speech. The 2.17B-parameter model is designed for on-device use and has downloadable BF16 and GGUF checkpoints.","releaseNote":null,"classifications":{"developmentType":"model","capability":["text generation","audio generation","analysis","extraction","translation"],"modality":["text","image","audio","video","multimodal"],"access":["API","web","open weights"],"deployment":["hosted","on-device"],"pricing":["unknown"],"targetUser":["consumer","developer","researcher"]},"citations":[{"url":"https://huggingface.co/LiquidAI/LFM3-Omni-2B","authority":"owner-controlled","publisher":"Liquid AI"},{"url":"https://huggingface.co/blog/LiquidAI/lfm3","authority":"other","publisher":"Hugging Face"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-21T16:58:40.000Z","effectiveAt":"2026-10-28T00:20:45.124Z","url":"https://wowsuchai.withglint.app/records/wsa_26mmc26i8b8p","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_26mmc26i8b8p","webUrl":"https://wowsuchai.withglint.app/records/wsa_26mmc26i8b8p"},{"recordId":"wsa_guk7j9k3f7bu","versionId":"ver_sjKL4aM6T4OT","eventType":"new_development","category":"model","title":"d1-Omni-2B","maker":"Liquid AI","summary":"d1-Omni-2B is a 2.17B-parameter decision model built on LFM3-Omni-2B that answers typed questions in one forward pass without generating tokens. It accepts text, images, video frames, and audio, and returns answers with probabilities.","releaseNote":null,"classifications":{"developmentType":"model","capability":["analysis","extraction"],"modality":["text","image","audio","video","multimodal"],"access":["API","SDK","open weights"],"deployment":["hosted","on-premises"],"pricing":["unknown"],"targetUser":["developer","enterprise"]},"citations":[{"url":"https://huggingface.co/LiquidAI/d1-Omni-2B","authority":"owner-controlled","publisher":"Liquid AI"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-24T19:30:58.000Z","effectiveAt":"2026-10-28T00:20:38.411Z","url":"https://wowsuchai.withglint.app/records/wsa_guk7j9k3f7bu","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_guk7j9k3f7bu","webUrl":"https://wowsuchai.withglint.app/records/wsa_guk7j9k3f7bu"},{"recordId":"wsa_ug7p3suzjfva","versionId":"ver_imfr4WplSHsI","eventType":"new_development","category":"model","title":"LFM3-1.7B","maker":"Liquid AI","summary":"LFM3-1.7B is a text model in Liquid AI’s LFM3 family, built for local agents and specialized applications. It has a 256K-token context window, is post-trained for agentic harnesses, and is available as open weights.","releaseNote":null,"classifications":{"developmentType":"model","capability":["text generation","analysis","code generation","agents","automation","fine-tuning"],"modality":["text","code"],"access":["API","web","open weights"],"deployment":["hosted","on-device","on-premises"],"pricing":["unknown"],"targetUser":["developer","enterprise","researcher"]},"citations":[{"url":"https://www.liquid.ai/blog/lfm3","authority":"owner-controlled","publisher":"Liquid AI"},{"url":"https://huggingface.co/LiquidAI/LFM3-1.7B","authority":"owner-controlled","publisher":"Liquid AI"},{"url":"https://huggingface.co/blog/LiquidAI/lfm3","authority":"other","publisher":"Hugging Face"},{"url":"https://github.com/ollama/ollama/releases/tag/v0.40.12","authority":"other","publisher":"GitHub watchlist"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-27T00:00:00.000Z","effectiveAt":"2026-10-28T00:20:36.467Z","url":"https://wowsuchai.withglint.app/records/wsa_ug7p3suzjfva","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_ug7p3suzjfva","webUrl":"https://wowsuchai.withglint.app/records/wsa_ug7p3suzjfva"},{"recordId":"wsa_9yeug8as48fe","versionId":"ver_koQnTV4cz4lk","eventType":"new_development","category":"model","title":"LFM3-1.7B-Base","maker":"Liquid AI","summary":"LFM3-1.7B-Base is a pre-trained, text-only 1.73B-parameter checkpoint in Liquid AI’s LFM3 family. It has a 256K-token context length and is intended as a starting point for substantial fine-tuning.","releaseNote":null,"classifications":{"developmentType":"model","capability":["text generation","fine-tuning"],"modality":["text"],"access":["open weights","API"],"deployment":["on-device"],"pricing":["unknown"],"targetUser":["developer","researcher"]},"citations":[{"url":"https://huggingface.co/LiquidAI/LFM3-1.7B-Base","authority":"owner-controlled","publisher":"Liquid AI"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-26T16:28:49.000Z","effectiveAt":"2026-10-28T00:20:34.357Z","url":"https://wowsuchai.withglint.app/records/wsa_9yeug8as48fe","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_9yeug8as48fe","webUrl":"https://wowsuchai.withglint.app/records/wsa_9yeug8as48fe"},{"recordId":"wsa_rv5dhcpijwn6","versionId":"ver_qUCqJX8lOU8j","eventType":"new_development","category":"model","title":"MiniCPM-OCR 1.0","maker":"OpenBMB","summary":"MiniCPM-OCR 1.0 is a 1.3B OCR-focused model for understanding documents, including text, tables, formulas, charts, seals, and structured information. It supports Chinese and English and is released with model weights, API documentation, SDK support, and deployment guides for several runtimes.","releaseNote":null,"classifications":{"developmentType":"model","capability":["extraction","analysis"],"modality":["image","text"],"access":["API","SDK","open weights"],"deployment":["on-premises"],"pricing":["unknown"],"targetUser":["developer","researcher","enterprise"]},"citations":[{"url":"https://huggingface.co/openbmb/MiniCPM-OCR-1.0","authority":"owner-controlled","publisher":"OpenBMB"},{"url":"https://github.com/OpenBMB/MiniCPM-OCR","authority":"owner-controlled","publisher":"GitHub new repositories"},{"url":"https://huggingface.co/openbmb/MiniCPM-OCR-1.0-DSpark","authority":"owner-controlled","publisher":"Hugging Face"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-21T15:51:40.000Z","effectiveAt":"2026-10-28T00:20:16.735Z","url":"https://wowsuchai.withglint.app/records/wsa_rv5dhcpijwn6","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_rv5dhcpijwn6","webUrl":"https://wowsuchai.withglint.app/records/wsa_rv5dhcpijwn6"},{"recordId":"wsa_v989cwdcn7ub","versionId":"ver_1D-6Wy6n88mB","eventType":"new_development","category":"model","title":"Hy-SimulMT-30B-A3B","maker":"Tencent Hunyuan","summary":"Hy-SimulMT-30B-A3B is a real-time simultaneous translation model for bidirectional Chinese–English speech translation. It continuously processes incoming audio and produces translated text and speech, with real-time voice cloning supported when a TTS service is configured.","releaseNote":null,"classifications":{"developmentType":"model","capability":["translation","audio generation"],"modality":["audio","text"],"access":["open weights"],"deployment":["on-premises"],"pricing":["unknown"],"targetUser":["developer"]},"citations":[{"url":"https://huggingface.co/tencent/Hy-SimulMT-30B-A3B","authority":"owner-controlled","publisher":"Tencent Hunyuan"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-22T07:29:17.000Z","effectiveAt":"2026-10-28T00:20:05.711Z","url":"https://wowsuchai.withglint.app/records/wsa_v989cwdcn7ub","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_v989cwdcn7ub","webUrl":"https://wowsuchai.withglint.app/records/wsa_v989cwdcn7ub"},{"recordId":"wsa_uvvu8wqvjfxj","versionId":"ver_PJGxmQ4LEIe6","eventType":"new_development","category":"model","title":"Hy-SimulMT-30B-A3B-FP8","maker":"Tencent Hunyuan","summary":"Hy-SimulMT-30B-A3B-FP8 is a real-time simultaneous translation model for bidirectional speech-to-speech translation between Chinese and English. It continuously processes incoming audio, produces translated text and speech, and supports real-time voice cloning; speech output and voice cloning require a configured TTS service.","releaseNote":null,"classifications":{"developmentType":"model","capability":["translation"],"modality":["audio","text"],"access":["open weights"],"deployment":["on-premises"],"pricing":["unknown"],"targetUser":["developer","researcher"]},"citations":[{"url":"https://huggingface.co/tencent/Hy-SimulMT-30B-A3B-FP8","authority":"owner-controlled","publisher":"Tencent Hunyuan"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-22T07:29:37.000Z","effectiveAt":"2026-10-28T00:19:48.890Z","url":"https://wowsuchai.withglint.app/records/wsa_uvvu8wqvjfxj","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_uvvu8wqvjfxj","webUrl":"https://wowsuchai.withglint.app/records/wsa_uvvu8wqvjfxj"},{"recordId":"wsa_e4xixqvv53t5","versionId":"ver_224PMg6jVPQS","eventType":"new_development","category":"model","title":"Unlimited OCR 2","maker":"Baidu ERNIE","summary":"Unlimited OCR 2 is an image-text model for parsing documents, including multi-page documents. It uses Sparse Reference Sliding Window Attention to support unlimited-horizon parsing within a finite context window.","releaseNote":null,"classifications":{"developmentType":"model","capability":["extraction"],"modality":["image","text","multimodal"],"access":["open weights","API","SDK"],"deployment":["on-premises","hosted"],"pricing":["unknown"],"targetUser":["developer","enterprise","researcher"]},"citations":[{"url":"https://huggingface.co/baidu/Unlimited-OCR-2","authority":"owner-controlled","publisher":"Hugging Face"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-26T03:57:46.000Z","effectiveAt":"2026-10-28T00:19:39.443Z","url":"https://wowsuchai.withglint.app/records/wsa_e4xixqvv53t5","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_e4xixqvv53t5","webUrl":"https://wowsuchai.withglint.app/records/wsa_e4xixqvv53t5"},{"recordId":"wsa_dpenhp6w8mt5","versionId":"ver_FrYEk5HMyv_G","eventType":"new_development","category":"developer platform","title":"Chaos Intelligence","maker":"Cattmonkey","summary":"Chaos Intelligence is a functional programming language that treats AI agents as callable values, with a CLI and web IDE. It supports building agent workflows, goal-driven bots, and interactive pages and reports.","releaseNote":null,"classifications":{"developmentType":"developer platform","capability":["agents","automation","code generation"],"modality":["text","code"],"access":["CLI","web","SDK"],"deployment":["on-premises"],"pricing":["unknown"],"targetUser":["developer"]},"citations":[{"url":"https://github.com/Cattmonkey/ChaosIntelligence","authority":"owner-controlled","publisher":"GitHub new repositories"}],"score":1,"confidence":"high","correctionState":"current","publishedAt":"2026-10-23T13:04:31.000Z","effectiveAt":"2026-10-28T00:21:14.933Z","url":"https://wowsuchai.withglint.app/records/wsa_dpenhp6w8mt5","apiUrl":"https://wowsuchai.withglint.app/api/v1/records/wsa_dpenhp6w8mt5","webUrl":"https://wowsuchai.withglint.app/records/wsa_dpenhp6w8mt5"}],"corrections":[],"notes":[],"writer":{"kind":"model","model":"openai/gpt-6-luna"},"publishedAt":"2026-10-29T00:36:23.000Z","updatedAt":"2026-10-29T00:36:23.000Z","url":"https://wowsuchai.withglint.app/briefs/daily/2026-10-28","apiUrl":"https://wowsuchai.withglint.app/api/v1/briefs/daily/2026-10-28"},"nextCursor":null,"license":{"catalog":{"name":"CC BY 4.0","spdx":"CC-BY-4.0","url":"https://creativecommons.org/licenses/by/4.0/","attribution":"wowsuchai","attributionUrl":"https://wowsuchai.withglint.app","covers":["catalog metadata","classifications","generated summaries","feed event metadata"]},"excluded":{"covers":["source quotations","linked source content"],"terms":"Not licensed by wowsuchai. Source quotations and linked source content keep their original rights."}}}