ReleasesToolsDiscounts

Model releases

Last scanned 26 September 2026, 18:54 UTC.

Every model this briefing has seen announced, newest first — one row per model rather than one per story, so the six outlets that cover a launch collapse into the release they are all covering. The date is the earliest any source in the sweep reported it, which for a model that shipped before this record opened, or that these sources came to late, is later than the day it launched. It is a record of coverage, not a vendor changelog.

Overview

  • 147models trackedsince 21 Jul 2026
  • 119in the last 30 days55 in the last 7
  • 42of those open-weightsdownloadable, not only served
  • 75labswith at least one release on record
  • 8seen, not launchedmet through a platform listing

6 Jul 2026this week

The newest from each lab

One row per organisation — its most recent release — with the labs that have shipped most in the last three months first. Pick the labs you follow and this table, and every release below it, narrow to them: the choice is remembered on this device and travels in the address bar, so a bookmark of your own labs is a link.

ReleasedModelLabAccessTypeReported by
2 days agoZhenwu V900Alibaba unveiled its self-developed Zhenwu V900 AI chip.Alibaba——DigiTimes
3 days agogemini-3.8-flash-lite-ttsGoogle released gemini-3.8-flash-lite-tts model.GoogleAPIaudioSimon Willison
2 days agoPUBG AllyIntroduces PUBG Ally, a conversational embodied agent for PUBG that reasons, acts autonomously, and plays as a voice-enabled teammate.UnknownresearchmultimodalarXiv cs.AI
4 days agoGPT-6 SolOpenAI releases GPT-6 Sol, cutting API costs by 50% or more.OpenAIAPItextVentureBeat AI+5
4 days agoClaude Opus 5.5Anthropic is launching Claude Opus 5.5, the first model in a new generation.AnthropicAPItextThe Decoder+6
3 days agoNVIDIA Nemotron 3 DiarizationNVIDIA Nemotron 3 Diarization enables real-time multi-speaker AI.Nvidiaopen weightsaudioHugging Face+1
3 days agoInternW0InternW0 is the first instantiation of the InternW physical world model series from Shanghai AI Laboratory.Shanghai AI Lab—worldarXiv cs.AI
4 days agoMiMo-V2.6-ProXiaomi released MiMo-V2.6-Pro as the top-performing open-weight model in the world.Xiaomiopen weightstextVentureBeat AI+2
5 days agoGrok 4.7Grok 4.7 is rolling out in GitHub Copilot for agentic coding and complex workflows.xAIin-appcodeGitHub Copilot+3
3 days agoME-VLMLi Auto released ME-VLM model.Li Auto—textPandaily
5 days agoMeta SAM 3.1Meta announced the SAM 3.1 model.Meta—imageTLDR AI
8 days agoGLM-5.3-FlashXZhipu AI launched GLM-5.3-FlashX with claimed ~200 tok/s inference on ~100,000 domestic accelerators.Z.aiopen weightstextPandaily
yesterdayGLiNER2.5-DecideFastino Labs released GLiNER2.5-Decide, a 340M-parameter open-weight decision model that runs on CPU.Fastinoopen weightstextMarktechpost
2 days agoJev-MobileJev-Mobile is introduced as an executor for mobile GUI agents.TypeSaferesearchmultimodalarXiv cs.AI
3 days agoFLUX 3 ActionBlack Forest Labs releases open-weight FLUX 3 Action robotics model.Black Forest Labsopen weightsroboticsVentureBeat AI+1
4 days agoSenseNova U1 ProSenseTime released the production version of its SenseNova U1 Pro image-creation model via its Raccoon app and API.SenseTimein-appimageTechNode+1
4 days agoLight-O1-PreviewLight Origins released Light-O1-Preview, a 6B Apache-2.0 model for robot whole-body trajectories.Light Originsopen weightsroboticsPandaily
5 days agoT1PrimeBot launched transformable T1 personal robot for sale.PrimeBot—roboticsPandaily
13 days agoIris-proThe AllSpark team released Iris-mini and Iris-pro, two open-source search agents built on Qwen models.AllSpark teamopen weightstextThe Decoder
2 weeks agoNorth Small TranslateCohere released North Small Translate, an open-weight Mixture-of-Experts model built for machine translation across 50 languages.Cohereopen weightstextMarktechpost
2 weeks agoFugu Ultra v2Sakana AI released Fugu Max and Fugu Ultra v2, two models for multi-agent orchestration.Sakana AIAPItextMarktechpost
2 weeks agoDPT-3 VerityLandingAI shipped Agentic Document Extraction Gen2, featuring the DPT-3 Verity model for word-level grounding.LandingAI—textMarktechpost
2 weeks agoDeepSeek-V4.1-FlashDeepSeek launched DeepSeek-V4.1-Flash with a 552-billion-parameter backbone, native vision, and a 1-million-token context window.DeepSeekAPImultimodalTechNode+5
4 weeks agoGigaPath-FlashMicrosoft Research presents GigaPath-Flash and GigaTIME-Flash pathology foundation models.MicrosoftresearchimageMicrosoft Research
todayAgent UltraExa released Agent Ultra, a high‑effort mode of its Exa Agent API for exhaustive list building.ExaAPItextMarktechpost
yesterdayGWM Worlds 2Runway announces GWM Worlds 2, a world model generating video and audio in real time.Runway—multimodalLatent Space
yesterdayAltar-1Aikido Security released Altar-1, an open-weight security model pruned from GLM-5.3.Aikido Securityopen weights—Marktechpost
2 days agoThinkingCap-Qwen3.8-27BBottleCap AI released ThinkingCap-Qwen3.8-27B, a fine-tune of Qwen3.8-27B with fewer thinking tokens.BottleCap AIopen weightstextMarktechpost
2 days agoSpark-ASR-2.0iFLYTEK shipped Spark-ASR-2.0 for noisy, dialect and Chinese-English code-switch recognition with an open API.iFLYTEKAPIspeechPandaily
2 days agoCLM-8BContrastive-LM released CLM-8B, an open System One model that scores agent actions faster than Jev.Contrastive-LMopen weightstextMarktechpost
3 days agotogether/Tev1-4B-experimentalTogether AI launches Jev-like classifier together/Tev1-4B-experimental on its serverless platform.Together AIAPItextTogether AI
3 days agoXing4.0-29B-A4BChina Telecom AI released Xing4.0-29B-A4B with open weights.China Telecom AIopen weightstextPandaily
3 days agoVoice of ReasonKyutai released two open-weight speech-to-speech models that solve spoken math with reinforcement learning.Kyutaiopen weightsspeechMarktechpost
5 days agoFuturist humanoidFaraday Future unveiled the all-new Futurist humanoid robot for sale.Faraday Future—roboticsDigiTimes
5 days agoAGILE 2.0AgiBot released AGILE 2.0, an end-to-end sense-control locomotion model for open-world environments.AgiBot—roboticsPandaily
6 days agoZGCM-1-7BZGCM-1-7B releases full weights, data, and code for math and agentic search.Zhongguancun Academyopen weightstextPandaily
6 days agoStep 5 PreviewStepFun announces Step 5 Preview, a 600B sparse MoE model with 1M context, weights opening Oct 15.StepFunopen weightstextPandaily+1
7 days agoSPARSEUPLinkup Research released SPARSEUP, an open-source sparse embedding model built on a 149M-parameter ModernBERT backbone.Linkup Researchopen weightsembeddingMarktechpost
8 days agojina-ocr-v1Jina AI has released jina-ocr-v1, a visual document parser that converts PDFs, scans, tables, charts and invoices into Markdown.Jina AI—imageMarktechpost
8 days agoZDTaichu5.0-9BTaichuAI open-sourced ZDTaichu5.0-9B, a ~9B spatial multimodal model.TaichuAIopen weightsmultimodalPandaily
8 days agoTernary Bonsai 2 27BPrismML released Ternary Bonsai 2 27B, a 5.9 GB Apache‑2.0 model supporting text and images.PrismMLopen weightsmultimodalMarktechpost
8 days agoNemotronLabs VoiceChatNemotronLabs VoiceChat is an open full-duplex speech-to-speech model with native tool-calling capabilities.NemotronLabsopen weightsspeecharXiv cs.AI
8 days agoNASA-IBM Lunar FoundationNASA and IBM released an open-source geospatial AI model for lunar applications.NASA-IBMopen weights—Hacker News AI
8 days agoDoubao-Seed-2.1-pro 0915ByteDance shipped Doubao-Seed-2.1-pro 0915, a multimodal coding model on Volcengine.ByteDance—multimodalPandaily
8 days agoCactus Needle 3Cactus Needle 3 model was announced, claiming to match DeepSeek V4 Flash performance.Cactus——Hacker News AI
8 days agoABot-Earth 0.7Amap launched ABot-Earth 0.7, a 3D-native urban world model generating cities from satellite or text input.Amap—worldPandaily
9 days agoPhysBrain 1.5DeepCybo open-sources PhysBrain 1.5, a 2B/8B physical foundation model.DeepCyboopen weightsmultimodalPandaily
9 days agoLimiX-2Stable AI and Tsinghua released LimiX-2, a 400M structured-data foundation model using Contextual Mechanism Networks.Stable AI——Pandaily
9 days agoHERON-CRASongyan Dynamics released the Scalabot HERON-CRA robot with 4D Context Expert memory and RL Engine.Songyan Dynamics—roboticsPandaily
9 days agoAPXInfInfinigence AI open-sources APXInf for embodied edge inference on Jetson Thor.Infinigence AIopen weights—Pandaily
10 days agoTabPFN-3.5Prior Labs releases TabPFN-3.5, a tabular foundation model that beats Otto's winning Kaggle solution with default settings.Prior Labs——Marktechpost
10 days agoSpecs IntelligenceSnap launches Specs Intelligence, a new AI assistant for iOS and Mac.Snapin-apptextThe Verge AI
10 days agoGLiFormerKnowledgator releases GLiFormer, a 575M-parameter encoder achieving 91.10 F1 on nested JSON extraction.Knowledgator—embeddingMarktechpost
10 days agoCog-WM 1.0EBKernel launched Cog-WM 1.0, a brain-inspired cognitive world model.EBKernel—worldPandaily
10 days agoCausiloNums AI released Causilo, a pretrained tabular foundation model for classification and regression.Nums AIopen weights—Marktechpost
12 days agoOM-1Reward AI has released OM-1, a robot policy trained solely on human demonstrations.Reward AI—roboticsMarktechpost
12 days agoMotus2Shengshu Technology releases Motus2, a self‑evolving world model for dexterous manipulation.ShengshuresearchworldPandaily
12 days agoKimi K2.8 PreviewMoonshot AI launches Kimi K2.8 Preview for Kimi Code and Work users via API.Moonshot AIAPIcodePandaily
2 weeks agoSWE-2Cognition released SWE-2, its most capable coding model to date, post-trained from Kimi K3.Cognitionopen weightscodeMarktechpost
2 weeks agoMiniCPM5-2BOpenBMB open-sourced MiniCPM5-2B, a 2.52B dense Apache-2.0 model.OpenBMBopen weightstextPandaily
3 weeks agoK2 HorizonIFM released K2 Horizon, a fleet of six Apache 2.0 models ranging from 0.9B to 375B.Institute of Foundation Modelsopen weightstextHacker News AI+1
4 weeks agoXenomiBeijing Guoguang Liangchao Technology introduced Xenomi, a family of large language models.Beijing Guoguang Liangchao Technology—textDigiTimes
4 weeks agoGradium AI Default TTSGradium AI releases a new default text-to-speech model with high pass rates.Gradium AIAPIspeechMarktechpost
4 weeks agoLux3DManycore Tech launched Lux3D, which generates 3D assets from text or images.Manycore TechAPIimageTechNode
4 weeks agoLing-3.0-flash-FinAnt Group launched the finance-tuned Ling-3.0-flash-Fin model, with weights scheduled for release next week.Ant Group—textTechNode
4 weeks agoHy4Tencent launched a preview version of its new flagship open-source model, Hy4.Tencentopen weightstextThe Information+1
4 weeks agoGranite 4.2IBM's new Granite 4.2 models ride the wave of interest in local LLMs.IBM—textArs Technica AI
5 weeks agoGEN-1.5Generalist AI released GEN-1.5, a robot foundation model that learns new physical tasks from a single demonstration.Generalist AI—roboticsMarktechpost
5 weeks agodots3-note previewXiaohongshu released the dots3-note preview, a MoE model with 280B total and 16B activated parameters.Xiaohongshuopen weightsmultimodalPandaily
5 weeks agoOmega-0Researchers released Omega-0, a latent-prediction world action model that lets a humanoid robot walk, look, and work.BAAIresearchroboticsPandaily
5 weeks agoS1-miniS1-mini is a 462 MB open-weights normalizer that sits after ASR, removing fillers and resolving self-corrections locally.Superwhisperopen weightstextMarktechpost
7 weeks agoPokee-Isaac 28BPokee AI released Pokee-Isaac 28B, a 28B text-only foundation model with a 10M-token context window.Pokee AI—textMarktechpost
8 weeks agoMiniMax-H3MiniMax released MiniMax-H3, a general-purpose, omni-modal generative system that accepts text, images, audio and video.MiniMaxopen weightsmultimodalThe Decoder+1
8 weeks agoInstella-MoE-16B-A3BAMD released Instella-MoE-16B-A3B, a fully open Mixture-of-Experts language model trained from scratch on Instinct GPUs.AMDopen weightstextMarktechpost
8 weeks agoInkling-SmallThinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model.Thinking Machines Labopen weightsmultimodalThe Decoder+1

Every release

Grouped by the month it was first reported. The link on each name opens the story this briefing read it in.

September 2026 · 111

ReleasedModelLabAccessTypeReported by
todayAgent UltraExa released Agent Ultra, a high‑effort mode of its Exa Agent API for exhaustive list building.ExaAPItextMarktechpost
yesterdayGWM Worlds 2Runway announces GWM Worlds 2, a world model generating video and audio in real time.Runway—multimodalLatent Space
yesterdayGLiNER2.5-DecideFastino Labs released GLiNER2.5-Decide, a 340M-parameter open-weight decision model that runs on CPU.Fastinoopen weightstextMarktechpost
yesterdayAltar-1Aikido Security released Altar-1, an open-weight security model pruned from GLM-5.3.Aikido Securityopen weights—Marktechpost
2 days agoZhenwu V900Alibaba unveiled its self-developed Zhenwu V900 AI chip.Alibaba——DigiTimes
2 days agoThinkingCap-Qwen3.8-27BBottleCap AI released ThinkingCap-Qwen3.8-27B, a fine-tune of Qwen3.8-27B with fewer thinking tokens.BottleCap AIopen weightstextMarktechpost
2 days agoSpark-ASR-2.0iFLYTEK shipped Spark-ASR-2.0 for noisy, dialect and Chinese-English code-switch recognition with an open API.iFLYTEKAPIspeechPandaily
2 days agoPUBG AllyIntroduces PUBG Ally, a conversational embodied agent for PUBG that reasons, acts autonomously, and plays as a voice-enabled teammate.UnknownresearchmultimodalarXiv cs.AI
2 days agoJev-MobileJev-Mobile is introduced as an executor for mobile GUI agents.TypeSaferesearchmultimodalarXiv cs.AI
2 days agoHEXISHEXIS compiles agent skills into extended finite state machines.UnknownresearchroboticsarXiv cs.AI
2 days agoGridSFMGridSFM is a framework for solving AC Optimal Power Flow.UnknownresearchsciencearXiv cs.LG
2 days agoGRASPGRASP is a strategy-aware, multi-stage planning framework for complex tasks.UnknownresearchtextarXiv cs.AI
2 days agoCLM-8BContrastive-LM released CLM-8B, an open System One model that scores agent actions faster than Jev.Contrastive-LMopen weightstextMarktechpost
3 days agotogether/Tev1-4B-experimentalTogether AI launches Jev-like classifier together/Tev1-4B-experimental on its serverless platform.Together AIAPItextTogether AI
3 days agogemini-3.8-flash-lite-ttsGoogle released gemini-3.8-flash-lite-tts model.GoogleAPIaudioSimon Willison
3 days agoXing4.0-29B-A4BChina Telecom AI released Xing4.0-29B-A4B with open weights.China Telecom AIopen weightstextPandaily
3 days agoVoice of ReasonKyutai released two open-weight speech-to-speech models that solve spoken math with reinforcement learning.Kyutaiopen weightsspeechMarktechpost
3 days agoQwen-Audio-3.1Alibaba debuted Qwen3.8-LiveTranslate and unveiled Qwen-Audio-3.1 family at Apsara 2026.AlibabaAPIaudioPandaily+1
3 days agoNVIDIA Nemotron 3 DiarizationNVIDIA Nemotron 3 Diarization enables real-time multi-speaker AI.Nvidiaopen weightsaudioHugging Face+1
3 days agoNV-Reason-CT Open 3D CT VLMNVIDIA introduces NV-Reason-CT Open 3D CT VLM for radiologist chain-of-thought reasoning.Nvidia—multimodalNVIDIA Developer
3 days agoME-VLMLi Auto released ME-VLM model.Li Auto—textPandaily
3 days agoME-U0Li Auto released ME-U0 model.Li Auto—textPandaily
3 days agoME-Brain-1.0Li Auto released ME-Brain-1.0 model.Li Auto—textPandaily
3 days agoLarge KnowledgeLarge Knowledge Model is introduced as a new model for scientific reasoning from papers.Unknown—sciencearXiv cs.CL
3 days agoInternW0InternW0 is the first instantiation of the InternW physical world model series from Shanghai AI Laboratory.Shanghai AI Lab—worldarXiv cs.AI
3 days agoGemini 3.8 Flash TTSGoogle launches Gemini 3.8 Flash TTS text-to-speech model.GoogleAPIaudioThe Decoder+2
3 days agoFlash-Lite TTSGoogle launches Flash-Lite TTS text-to-speech model.GoogleAPIaudioThe Decoder+1
3 days agoFLUX 3 ActionBlack Forest Labs releases open-weight FLUX 3 Action robotics model.Black Forest Labsopen weightsroboticsVentureBeat AI+1
4 days agoSenseNova U1 ProSenseTime released the production version of its SenseNova U1 Pro image-creation model via its Raccoon app and API.SenseTimein-appimageTechNode+1
4 days agoMiMo-V2.6-ProXiaomi released MiMo-V2.6-Pro as the top-performing open-weight model in the world.Xiaomiopen weightstextVentureBeat AI+2
4 days agoMiMo-V2.6-FlashXiaomi released MiMo-V2.6-Flash alongside the Pro version as a cheaper option.Xiaomiopen weightstextVentureBeat AI+2
4 days agoMiMo-V2.6Xiaomi released and open-sourced its MiMo-V2.6 series.Xiaomiopen weightstextDigiTimes
4 days agoLight-O1-PreviewLight Origins released Light-O1-Preview, a 6B Apache-2.0 model for robot whole-body trajectories.Light Originsopen weightsroboticsPandaily
4 days agoIntern-S2-397BShanghai AI Laboratory released Intern-S2-397B, an open-weight science multimodal model with a pluggable Memory Decoder.Shanghai AI Labopen weightsmultimodalPandaily
4 days agoGPT-6 SolOpenAI releases GPT-6 Sol, cutting API costs by 50% or more.OpenAIAPItextVentureBeat AI+5
4 days agoGPT-6 LunaOpenAI releases GPT-6 Luna, cutting API costs by 50% or more.OpenAIAPItextVentureBeat AI+5
4 days agoClaude Opus 5.5Anthropic is launching Claude Opus 5.5, the first model in a new generation.AnthropicAPItextThe Decoder+6
5 days agot0-betaThe authors released t0-beta (256M) open‑weight time‑series model for forecasting.Unknownopen weightssciencearXiv cs.LG
5 days agot0-alphaThe authors released t0-alpha (102M) open‑weight time‑series model for forecasting.Unknownopen weightssciencearXiv cs.LG
5 days agoT1PrimeBot launched transformable T1 personal robot for sale.PrimeBot—roboticsPandaily
5 days agoQ1PrimeBot launched foldable Q1 personal robot for sale.PrimeBot—roboticsPandaily
5 days agoNVIDIA Earth-2NVIDIA Earth-2 helps weather-sensitive industries make timely decisions using local observations.Nvidia—imageNVIDIA Developer
5 days agoMeta SAM 3.1Meta announced the SAM 3.1 model.Meta—imageTLDR AI
5 days agoGrok 4.7Grok 4.7 is rolling out in GitHub Copilot for agentic coding and complex workflows.xAIin-appcodeGitHub Copilot+3
5 days agoFuturist humanoidFaraday Future unveiled the all-new Futurist humanoid robot for sale.Faraday Future—roboticsDigiTimes
5 days agoAGILE 2.0AgiBot released AGILE 2.0, an end-to-end sense-control locomotion model for open-world environments.AgiBot—roboticsPandaily
6 days agoZGCM-1-7BZGCM-1-7B releases full weights, data, and code for math and agentic search.Zhongguancun Academyopen weightstextPandaily
6 days agoStep 5 PreviewStepFun announces Step 5 Preview, a 600B sparse MoE model with 1M context, weights opening Oct 15.StepFunopen weightstextPandaily+1
6 days agoQwen3.8-LiveTranslateAlibaba Qwen team released Qwen3.8-LiveTranslate, a real-time interpretation model.Alibaba—speechMarktechpost+1
6 days agoQwen-Image-2.1Qwen-Image-2.1 is introduced as a compact, efficient, unified image creation model.Alibabaopen weightsimageHacker News AI+4
6 days agoQwenTry Qwen at qwen.aiAlibaba—textQwen
7 days agoSenseNova U1.5SenseTime releases SenseNova U1.5, an 8B Mixture-of-Transformers vision model with open training code.SenseTime—imagePandaily
7 days agoSPARSEUPLinkup Research released SPARSEUP, an open-source sparse embedding model built on a 149M-parameter ModernBERT backbone.Linkup Researchopen weightsembeddingMarktechpost
7 days agoGrok Voice Transcribe 2.0SpaceXAI released Grok Voice Transcribe 2.0, a speech-to-text API claiming 2x accuracy over 1.0.xAIAPIspeechMarktechpost
7 days agoDAMO RADARAlibaba DAMO Academy open-sources the DAMO RADAR abdominal CT model.Alibabaopen weightsimagePandaily
8 days agojina-ocr-v1Jina AI has released jina-ocr-v1, a visual document parser that converts PDFs, scans, tables, charts and invoices into Markdown.Jina AI—imageMarktechpost
8 days agoZDTaichu5.0-9BTaichuAI open-sourced ZDTaichu5.0-9B, a ~9B spatial multimodal model.TaichuAIopen weightsmultimodalPandaily
8 days agoTernary Bonsai 2 27BPrismML released Ternary Bonsai 2 27B, a 5.9 GB Apache‑2.0 model supporting text and images.PrismMLopen weightsmultimodalMarktechpost
8 days agoSamsoneA family of open small audio language models for on-device inference.Unknown—audioarXiv cs.AI
8 days agoNemotronLabs VoiceChatNemotronLabs VoiceChat is an open full-duplex speech-to-speech model with native tool-calling capabilities.NemotronLabsopen weightsspeecharXiv cs.AI
8 days agoNASA-IBM Lunar FoundationNASA and IBM released an open-source geospatial AI model for lunar applications.NASA-IBMopen weights—Hacker News AI
8 days agoGLM-5.3-FlashXZhipu AI launched GLM-5.3-FlashX with claimed ~200 tok/s inference on ~100,000 domestic accelerators.Z.aiopen weightstextPandaily
8 days agoDoubao-Seed-2.1-pro 0915ByteDance shipped Doubao-Seed-2.1-pro 0915, a multimodal coding model on Volcengine.ByteDance—multimodalPandaily
8 days agoDamo Academy Medical AIAlibaba's Damo Academy open-sourced a model detecting 150 abdominal conditions from CT scans.Alibabaopen weightsimageSCMP Tech
8 days agoCactus Needle 3Cactus Needle 3 model was announced, claiming to match DeepSeek V4 Flash performance.Cactus——Hacker News AI
8 days agoAstra for LawOpenAI introduced Astra for Law, a version of its GPT-6 Astra model built for legal work.OpenAI—textThe Decoder
8 days agoABot-Earth 0.7Amap launched ABot-Earth 0.7, a 3D-native urban world model generating cities from satellite or text input.Amap—worldPandaily
9 days agoQwen 3.8 Omni FlashAlibaba releases Qwen 3.8 Omni Flash model.AlibabaAPImultimodalHacker News AI+3
9 days agoPhysBrain 1.5DeepCybo open-sources PhysBrain 1.5, a 2B/8B physical foundation model.DeepCyboopen weightsmultimodalPandaily
9 days agoLimiX-2Stable AI and Tsinghua released LimiX-2, a 400M structured-data foundation model using Contextual Mechanism Networks.Stable AI——Pandaily
9 days agoLightNav-0Light Origins open-sourced LightNav-0, a Qwen3-VL-4B navigation model trained via Real2Sim2Real on 2,000+ scenes.Light Originsopen weightsroboticsPandaily
9 days agoHERON-CRASongyan Dynamics released the Scalabot HERON-CRA robot with 4D Context Expert memory and RL Engine.Songyan Dynamics—roboticsPandaily
9 days agoAPXInfInfinigence AI open-sources APXInf for embodied edge inference on Jetson Thor.Infinigence AIopen weights—Pandaily
10 days agoXiaomi-Robotics-U0Xiaomi open-sourced Xiaomi-Robotics-U0, an embodied world foundation model with 4B and 38B weights.Xiaomiopen weightsworldPandaily
10 days agoTabPFN-3.5Prior Labs releases TabPFN-3.5, a tabular foundation model that beats Otto's winning Kaggle solution with default settings.Prior Labs——Marktechpost
10 days agoSpecs IntelligenceSnap launches Specs Intelligence, a new AI assistant for iOS and Mac.Snapin-apptextThe Verge AI
10 days agoJevTypeSafe released Jev, a System One model that decides, classifies, routes, and scores.TypeSafeAPItextLatent Space+5
10 days agoGLiFormerKnowledgator releases GLiFormer, a 575M-parameter encoder achieving 91.10 F1 on nested JSON extraction.Knowledgator—embeddingMarktechpost
10 days agoCog-WM 1.0EBKernel launched Cog-WM 1.0, a brain-inspired cognitive world model.EBKernel—worldPandaily
10 days agoCausiloNums AI released Causilo, a pretrained tabular foundation model for classification and regression.Nums AIopen weights—Marktechpost
10 days agoAtria Dawn PreviewShanghai Artificial Intelligence Laboratory open-sourced Atria Dawn Preview, a 744B MoE agentic model.Shanghai AI Labopen weightstextPandaily
11 days agoGemini 3.8 Live Extended ThinkingGoogle DeepMind introduces Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking.GoogleAPImultimodalGoogle DeepMind+2
11 days agoGemini 3.8 LiveGoogle DeepMind introduces Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking.GoogleAPImultimodalGoogle DeepMind+2
12 days agoOM-1Reward AI has released OM-1, a robot policy trained solely on human demonstrations.Reward AI—roboticsMarktechpost
12 days agoNCP-ArchPreviewShanghai AI Lab and SJTU open-sourced NCP-ArchPreview, an 8.9B latent-space model.Shanghai AI Labopen weightstextPandaily
12 days agoMotus2Shengshu Technology releases Motus2, a self‑evolving world model for dexterous manipulation.ShengshuresearchworldPandaily
12 days agoKimi K2.8 PreviewMoonshot AI launches Kimi K2.8 Preview for Kimi Code and Work users via API.Moonshot AIAPIcodePandaily
12 days agoClaude for Financial AdvisorsAnthropic introduces Claude for Financial Advisors.Anthropic—textAnthropic
13 days agoIris-proThe AllSpark team released Iris-mini and Iris-pro, two open-source search agents built on Qwen models.AllSpark teamopen weightstextThe Decoder
13 days agoIris-miniThe AllSpark team released Iris-mini and Iris-pro, two open-source search agents built on Qwen models.AllSpark teamopen weightstextThe Decoder
2 weeks agoSWE-2Cognition released SWE-2, its most capable coding model to date, post-trained from Kimi K3.Cognitionopen weightscodeMarktechpost
2 weeks agoNorth Small TranslateCohere released North Small Translate, an open-weight Mixture-of-Experts model built for machine translation across 50 languages.Cohereopen weightstextMarktechpost
2 weeks agoMiniCPM5-2BOpenBMB open-sourced MiniCPM5-2B, a 2.52B dense Apache-2.0 model.OpenBMBopen weightstextPandaily
2 weeks agoFugu Ultra v2Sakana AI released Fugu Max and Fugu Ultra v2, two models for multi-agent orchestration.Sakana AIAPItextMarktechpost
2 weeks agoFugu MaxSakana AI released Fugu Max and Fugu Ultra v2, two models for multi-agent orchestration.Sakana AIAPItextMarktechpost
2 weeks agoGPT‑Live‑1GPT‑Live‑1 brings natural, full-duplex voice conversations to the API with stronger instruction following.OpenAIAPIspeechOpenAI
2 weeks agoDPT-3 VerityLandingAI shipped Agentic Document Extraction Gen2, featuring the DPT-3 Verity model for word-level grounding.LandingAI—textMarktechpost
2 weeks agoDPT-3 ProLandingAI shipped Agentic Document Extraction Gen2, featuring the DPT-3 Pro model for line-level grounding.LandingAI—textMarktechpost
2 weeks agoDeepSeek-V4.1-FlashDeepSeek launched DeepSeek-V4.1-Flash with a 552-billion-parameter backbone, native vision, and a 1-million-token context window.DeepSeekAPImultimodalTechNode+5
3 weeks agoClaude Sonnet 5Anthropic announces Claude Sonnet 5.Anthropic—textAnthropic
3 weeks agoChatGPT Images 2.5ChatGPT Images 2.5 helps turn ideas and sketches into personalized images.OpenAIin-appimageOpenAI
3 weeks agoQwen-Drive-1.0-4BAlibaba’s Qwen team has released Qwen-Drive-1.0-4B, an open-source model for autonomous driving.Alibabaopen weightsmultimodalTechNode+1
3 weeks agoLyria 3.5Google has released its Lyria 3.5 music model in the Gemini app and via API.GoogleAPIaudioGoogle (Gemini)+1
3 weeks agoWeatherNext 3Google DeepMind introduces WeatherNext 3, its most advanced global weather AI model.Google—scienceGoogle DeepMind
3 weeks agoMuse Spark 1.3Meta AI Released Muse Spark 1.3, an agentic coding model with fewer tool calls.Meta—codeMarktechpost
3 weeks agoK2 HorizonIFM released K2 Horizon, a fleet of six Apache 2.0 models ranging from 0.9B to 375B.Institute of Foundation Modelsopen weightstextHacker News AI+1
3 weeks agoGPT-6 AstraOpenAI introduced GPT-6 Astra, its most capable model for business with advanced reasoning and computer use.OpenAIAPImultimodalHacker News AI+5
3 weeks agoGemini 3.8 FlashGoogle announced two versions of a new 3.8 Flash model.GoogleAPImultimodalGoogle DeepMind+2
4 weeks agoXenomiBeijing Guoguang Liangchao Technology introduced Xenomi, a family of large language models.Beijing Guoguang Liangchao Technology—textDigiTimes
4 weeks agoGradium AI Default TTSGradium AI releases a new default text-to-speech model with high pass rates.Gradium AIAPIspeechMarktechpost
4 weeks agoClaude Fable 5.1Anthropic launches Claude Fable 5.1 and says it is up to 45 percent cheaper.AnthropicAPItextThe Verge AI+2

August 2026 · 33

ReleasedModelLabAccessTypeReported by
4 weeks agoTimesFM-3Google Research released TimesFM-3, a 330-million-parameter forecasting model that analyzes time series data.GoogleresearchscienceGoogle Research+1
4 weeks agoGigaPath-FlashMicrosoft Research presents GigaPath-Flash and GigaTIME-Flash pathology foundation models.MicrosoftresearchimageMicrosoft Research
4 weeks agoLux3DManycore Tech launched Lux3D, which generates 3D assets from text or images.Manycore TechAPIimageTechNode
4 weeks agoLing-3.0-flash-FinAnt Group launched the finance-tuned Ling-3.0-flash-Fin model, with weights scheduled for release next week.Ant Group—textTechNode
4 weeks agoHy4Tencent launched a preview version of its new flagship open-source model, Hy4.Tencentopen weightstextThe Information+1
4 weeks agoParse 5Cohere released Parse 5, a 2.3B-parameter vision language model that converts documents into Markdown.CohereAPImultimodalMarktechpost
4 weeks agoGemini Omni 1.1 FlashGoogle DeepMind released Gemini Omni 1.1 Flash, allowing users to build with more control.Google—multimodalGoogle DeepMind
4 weeks agoGLM-5.3 FlashZ.ai’s new open-source model, GLM-5.3 Flash, was officially released earlier this week.Z.aiopen weightstextSCMP Tech+1
4 weeks agoQwen3.8-Flash-NextQwen released Qwen3.8-Flash-Next, an open weights multimodal MoE model.Alibabaopen weightsmultimodalSimon Willison
4 weeks agoGrok TtsxAI released Grok Tts.xAI—speechxAI
4 weeks agoGranite 4.2IBM's new Granite 4.2 models ride the wave of interest in local LLMs.IBM—textArs Technica AI
4 weeks agoGlucoFMGoogle Research released GlucoFM, a foundation model for continuous glucose monitoring.GoogleresearchscienceGoogle Research
4 weeks agoGemini 3.5 TranscribeGoogle announced Gemini 3.5 Transcribe for AI-powered speech-to-text.Googlein-appspeechArs Technica AI+1
5 weeks agoWan3.0Alibaba's video generation model Wan3.0 creates clips up to 30 seconds long from text, PDFs, and PowerPoint files.AlibabaAPIvideoThe Decoder
5 weeks agoGLiNER2.5Fastino released GLiNER2.5, a boundary-prediction architecture for information extraction with three Apache 2.0 checkpoints.Fastinoopen weightstextMarktechpost
5 weeks agoGEN-1.5Generalist AI released GEN-1.5, a robot foundation model that learns new physical tasks from a single demonstration.Generalist AI—roboticsMarktechpost
5 weeks agodots3-note previewXiaohongshu released the dots3-note preview, a MoE model with 280B total and 16B activated parameters.Xiaohongshuopen weightsmultimodalPandaily
5 weeks agoWan-Animate-2Alibaba's Tongyi Wanxiang team open-sourced Wan-Animate-2, an end-to-end character animation framework running at 24 fps.Alibabaopen weightsvideoPandaily
5 weeks agoV4-Flash-Vision-ExpDeepseek released V4-Flash-Vision-Exp, an experimental multimodal model that adds image understanding to V4-Flash's text capabilities.DeepSeekAPImultimodalThe Decoder+1
5 weeks agoQwen-UI-AgentAlibaba Qwen released Qwen-UI-Agent, a GUI agent foundation model for real devices covering phones, desktops, and web.Alibaba—multimodalPandaily
5 weeks agoOmega-0Researchers released Omega-0, a latent-prediction world action model that lets a humanoid robot walk, look, and work.BAAIresearchroboticsPandaily
5 weeks agoS1-miniS1-mini is a 462 MB open-weights normalizer that sits after ASR, removing fillers and resolving self-corrections locally.Superwhisperopen weightstextMarktechpost
6 weeks agoQwen3.8-27BQwen3.8-27B was released August 14, passing a million downloads in two days and hitting Hugging Face's global trend list.Alibabaopen weightstextSimon Willison+2
6 weeks agoGLM-5.3GLM-5.3 is now open-weight.Z.aiopen weightstextVentureBeat AI+1
7 weeks agoCARE-XMicrosoft Research introduces CARE-X, a unified approach for clinically useful radiology VLMs.MicrosoftresearchimageMicrosoft Research
7 weeks agoNemotronLabs VoiceChat 11BNVIDIA releases NemotronLabs VoiceChat 11B, an open full-duplex speech-to-speech model with live tool calling.Nvidiaopen weightsspeechMarktechpost
7 weeks agoPokee-Isaac 28BPokee AI released Pokee-Isaac 28B, a 28B text-only foundation model with a 10M-token context window.Pokee AI—textMarktechpost
7 weeks agoImagine Image 2.0xAI has released Imagine Image 2.0 as a new image generator for Grok.xAIin-appimageThe Decoder
7 weeks agoMuse Spark 1.2Meta shipped Muse Spark 1.2, a coding-focused update to Muse Spark 1.1.Metain-appcodeSimon Willison+1
7 weeks agoFLUX 3 VideoBlack Forest Labs has launched FLUX 3 Video, which generates Full HD clips up to 20 seconds long.Black Forest LabsAPIvideoThe Decoder
8 weeks agoQwen3.8-MaxAlibaba released Qwen3.8-Max, claiming performance rivaling the best systems from US frontier labs.Alibaba—textThe Verge AI+1
8 weeks agoMiniMax-H3MiniMax released MiniMax-H3, a general-purpose, omni-modal generative system that accepts text, images, audio and video.MiniMaxopen weightsmultimodalThe Decoder+1
8 weeks agoInstella-MoE-16B-A3BAMD released Instella-MoE-16B-A3B, a fully open Mixture-of-Experts language model trained from scratch on Instinct GPUs.AMDopen weightstextMarktechpost

July 2026 · 3

ReleasedModelLabAccessTypeReported by
8 weeks agoInkling-SmallThinking Machines Lab Releases Inkling-Small: A 276B Total, 12B Active Open Weights Multimodal MoE Model.Thinking Machines Labopen weightsmultimodalThe Decoder+1
2 months agoClaude Opus 5Anthropic is introducing Claude Opus 5.Anthropic—textAnthropic+1
2 months agoGemini 3.6 FlashGoogle DeepMind is introducing new Gemini models, including Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber.Google—textGoogle DeepMind

Seen, but not a launch

Models this briefing met through a distribution story — a cloud, an IDE or an inference platform adding a model that already existed — rather than through its announcement. The date is when that story ran, which is not when the model shipped, so they are kept out of the release list above and out of its counts. They are here because the sighting is real: this is the row for a model whose launch no source on the list covered.

ReleasedModelLabAccessTypeReported by
2 days agolocal languageA frozen local language model with four billion parameters is used for candidate generation.UnknownresearchtextarXiv cs.AI
5 days agoYuanzai robot dogDongfeng Motor announced the Yuanzai robot dog will soon be available at dealerships.Dongfeng Motor—roboticsPandaily
5 days agoGrok 4.6xAI's Grok 4.6 is now available in Amazon Bedrock for long-running agents and coding.xAIAPItextAWS Machine Learning
8 days agoKimi K3Kimi K3 from Moonshot AI is now available on Amazon Bedrock as an open-weight model for coding and knowledge work.Moonshot AIopen weightsmultimodalAWS Machine Learning+1
2 weeks agoGLM 5.2GLM 5.2 is now available on CoreWeave inference as an API.Z.aiAPItextCoreWeave
5 weeks agoGPT-5.6GPT-5.6 is now available in Kiro, helping developers plan, build, review, and test software.OpenAIAPIcodeOpenAI
6 weeks agoDeepSeek V4 Pro 0813The latest DeepSeek Pro model is now available via API only on OpenRouter.DeepSeekAPItextSimon Willison
3 months agoKimi K2.7 CodeCoreWeave Inference offers the newly-launched Kimi K2.7 Code on serverless inference.Moonshot AIAPIcodeCoreWeave

How this page is built

Each sweep reads the same candidate pool the briefing is selected from, and a model decides which stories announce a named release rather than merely discussing one — a distinction a keyword cannot make, since “Anthropic launches Claude Fable 5.1” and “Claude Fable 5.1 made me a nice pelican” are one sentence to a pattern. What it finds is folded into a cumulative record, keeping the earliest date and counting the outlets. When no model is reachable a stricter pattern stands in, and rows it produced are marked unconfirmed until a later run can check them. Nothing here is a vendor announcement: it is what the sources on the source list reported, and a release no source covered is a release this page does not know about.

Two things it refuses to treat as launches. A story that a model is now available on some cloud, IDE or inference platform is a distribution note about a model that already existed, and it is filed below the releases rather than dated as one. And a version cannot be newer than a higher version of itself: where a lab is already on record at 5.3, a later story about 5.2 is somebody catching up, not a release, and it is moved out of the list. The lab a model is filed under comes from the family its name belongs to before it comes from whatever the story called it — GLM is Z.ai's whoever is serving it.

2 digest-only sightings are held off this list until a second outlet or firmer detail confirms them.

One bridge sighting is held off this list until a second outlet or firmer detail confirms it.