#multimodalai — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #multimodalai, aggregated by home.social.
-
RT @Alibaba_Qwen: 📢Treffen Sie Qwen3.8-Max — unser leistungsfähigstes Modell bis heute. Nächste Woche werden die offenen Gewichte von Qwen3.8-Max veröffentlicht, und Qwen3.8-27B wird ebenfalls als Open-Weights verfügbar sein! 🎉 Qwen3.8-Max setzt einen neuen Maßstab für Coding und Zusammenarbeit bei 2,4 Billionen Parametern: - Autonomes Coding: 10+ Tage selbstentwickelter Programmierung, vom leeren Ordner bis zur Produktion ohne Anleitung, vollständiger Projektverlauf auf GitHub: https://github.com/qwen-code-dev-bot/oh-my-cli - Echte Arbeit, echte Ergebnisse: Produktionsreife Ergebnisse in Hunderten von Berufen. - Langfristige Beherrschung: Systemweites autonomes Planen mit geschlossenen adaptiven Lernschleifen, die 500+ Durchgänge der Chipdesign-Optimierung und 365 Tage E-Commerce-Strategie vorantreiben. - Native multimodale Intelligenz: Vision ist nicht nur Eingabe — sie ist ein kontinuierlicher Feedback-Loop für Planung, Ausführung und Selbstkorrektur. 💰Preise: Input: $2,0 / M Tokens Output: $6,0 / M Tokens Implicit Caching: $0,25 / M Tokens Beginnen Sie mit Qwen3.8-Max zu bauen! 🚀 📖 Blog: https://qwen.ai/blog?id=qwen3.8 ✅ Qwen Studio: https://chat.qwen.ai/?models=qwen3.8-max ⚡ API: https://www.qwencloud.com/models/qwen3.8-max
mehr auf Arint.info
#AIModel #Coding #MachineLearning #MultimodalAI #OpenWeights #Qwen3 #arint_info
-
RT @ComfyUI: MiniMax H3 ist jetzt über Partner Nodes in ComfyUI verfügbar. → Multimodale Ein- und Ausgabe: T2V, Erster/Letzter Frame, Omni Reference → Native Stereo-Audio auf jedem Clip → Bis zu 2K, 5-15 Sekunden bei 24 FPS → Charaktere, Szenen, Dialoge und Stimme können direkt vor Ort bearbeitet werden. API-Zugriff ist ab heute verfügbar. Native Open-Weight-Unterstützung folgt in Kürze. Video
mehr auf Arint.info
#APIAccess #ComfyUI #MiniMaxH3 #MultimodalAI #OpenWeight #VideoGeneration #arint_info
-
RT @ComfyUI: MiniMax H3 ist jetzt über Partner Nodes in ComfyUI verfügbar. → Multimodale Ein-/Ausgabe: Text-zu-Video, Erster/Letzter Frame, Omni-Referenz → Native Stereo-Audio auf jedem Clip → Bis zu 2K, 5-15 Sekunden bei 24 FPS → Charaktere, Szenen, Dialoge und Stimme können direkt vor Ort bearbeitet werden. API-Zugriff ist ab heute verfügbar. Native Open-Weight-Unterstützung folgt in Kürze. Video
mehr auf Arint.info
#APIAccess #ComfyUI #MiniMaxH3 #MultimodalAI #OpenWeight #VideoGeneration #arint_info
-
RT @MiniMax_AI: MiniMax H3: Omni-Referenz, kommerzielle Generierungsqualität, unschlagbare Kosteneffizienz, offene Gewichte. MiniMax H3: Ein offenes Modell, das die Grenzen zwischen Aufgaben und Modalitäten überwindet. Heute starten wir MiniMax H3, ein multimodales Generierungsmodell für allgemeine Zwecke. H3 versteht den kontextuellen Zusammenhang von Text, Bildern, Video und Audio und generiert Videos mit nativem Stereo-Sound, bis zu
mehr auf Arint.info
#GenerativeAI #Kosteneffizienz #MiniMaxH3 #MultimodalAI #OpenSource #VideoGeneration #arint_info
-
RT @MiniMax_AI: MiniMax H3: Omni-Referenz, kommerzielle Generierungsqualität, unschlagbare Kosteneffizienz, offene Gewichte. MiniMax H3: Ein offenes Modell, das die Grenzen zwischen Aufgaben und Modalitäten überwindet. Heute starten wir MiniMax H3, ein multimodales Generierungsmodell für allgemeine Zwecke. H3 versteht den kontextuellen Zusammenhang von Text, Bildern, Video und Audio und generiert Videos mit nativem Stereo-Sound, bis zu
mehr auf Arint.info
#GenerativeAI #Kosteneffizienz #MiniMaxH3 #MultimodalAI #OpenSource #VideoGeneration #arint_info
-
Black Forest Labs has launched FLUX 3 with support for 20-second video generation with synchronized audio.
#AI #FLUX #BlackForestLabs #AIVideoGeneration #MultimodalAI #GenAI #AIModels #AIImageGeneration #TextToVideo
-
Black Forest Labs has launched FLUX 3 with support for 20-second video generation with synchronized audio.
#AI #FLUX #BlackForestLabs #AIVideoGeneration #MultimodalAI #GenAI #AIModels #AIImageGeneration #TextToVideo
-
https://winbuzzer.com/2026/07/20/google-vids-rolls-out-personal-avatars-and-gemini-omni-xcxwbn/
Google Vids has started rolling out Personal AI Avatar creation and Gemini Omni for eligible customers, with account-bound likenesses & invisible SynthID watermarks.
#AI #GoogleVids #GeminiOmni #AIAvatars #Google #GoogleGemini #GoogleAI #GenAI #MultimodalAI #AIVideo
-
https://winbuzzer.com/2026/07/20/google-vids-rolls-out-personal-avatars-and-gemini-omni-xcxwbn/
Google Vids has started rolling out Personal AI Avatar creation and Gemini Omni for eligible customers, with account-bound likenesses & invisible SynthID watermarks.
#AI #GoogleVids #GeminiOmni #AIAvatars #Google #GoogleGemini #GoogleAI #GenAI #MultimodalAI #AIVideo
-
https://winbuzzer.com/2026/07/17/moonshot-ai-unveils-28t-parameter-kimi-k3-ai-model-xcxwbn/
Moonshot AI has launched its 2.8-trillion-parameter Kimi K3 model, but a high hallucination rate might tempers its frontier-model pitch.
#AI #MoonshotAI #KimiK3 #AIModels #MultimodalAI #MixtureOfExperts #AIReasoningModels #AIBenchmarks #ChinaAI
-
https://winbuzzer.com/2026/07/17/moonshot-ai-unveils-28t-parameter-kimi-k3-ai-model-xcxwbn/
Moonshot AI has launched its 2.8-trillion-parameter Kimi K3 model, but a high hallucination rate might tempers its frontier-model pitch.
#AI #MoonshotAI #KimiK3 #AIModels #MultimodalAI #MixtureOfExperts #AIReasoningModels #AIBenchmarks #ChinaAI
-
Google redesigned Workspace icons for people. Embedding analysis suggests the new designs are also easier for vision models to distinguish. https://hackernoon.com/did-googles-workspace-redesign-make-its-icons-easier-for-ai-to-see #multimodalai
-
Google redesigned Workspace icons for people. Embedding analysis suggests the new designs are also easier for vision models to distinguish. https://hackernoon.com/did-googles-workspace-redesign-make-its-icons-easier-for-ai-to-see #multimodalai
-
https://winbuzzer.com/2026/06/06/alibaba-pitches-qwen37-plus-as-a-computer-use-ai-agent-xcxwbn/
Alibaba's new Qwen3.7-Plus model targets screen, coding, and cloud-console automation as computer-use AI pushes beyond browsers into app and terminal tasks.
#AI #Qwen37Plus #Qwen3 #Alibaba #Qwen #AIAutomation #AIAgents #MultimodalAI #AICoding #AIModels
-
https://winbuzzer.com/2026/06/06/alibaba-pitches-qwen37-plus-as-a-computer-use-ai-agent-xcxwbn/
Alibaba's new Qwen3.7-Plus model targets screen, coding, and cloud-console automation as computer-use AI pushes beyond browsers into app and terminal tasks.
#AI #Qwen37Plus #Qwen3 #Alibaba #Qwen #AIAutomation #AIAgents #MultimodalAI #AICoding #AIModels
-
https://winbuzzer.com/2026/06/04/google-gemma-4-12b-targets-local-ai-agents-on-laptops-xcxwbn/
Google has released Gemma 4 12B, a local multimodal AI model for laptops that handles audio, images, code, and tool calls with 16GB memory locally.
#AI #Gemma4E4B #Gemma4 #GoogleGemma #Gemma #Google #GoogleAI #GoogleDeepMind #AIModels #MultimodalAI #OpenSourceAI #OnDeviceAI #AIAgents #AgenticAI
-
https://winbuzzer.com/2026/06/04/google-gemma-4-12b-targets-local-ai-agents-on-laptops-xcxwbn/
Google has released Gemma 4 12B, a local multimodal AI model for laptops that handles audio, images, code, and tool calls with 16GB memory locally.
#AI #Gemma4E4B #Gemma4 #GoogleGemma #Gemma #Google #GoogleAI #GoogleDeepMind #AIModels #MultimodalAI #OpenSourceAI #OnDeviceAI #AIAgents #AgenticAI
-
Google wants your next AI agent running locally on a 16GB laptop
https://fed.brid.gy/r/https://nerds.xyz/2026/06/google-gemma-4-12b-local-ai/
-
Google wants your next AI agent running locally on a 16GB laptop
https://web.brid.gy/r/https://nerds.xyz/2026/06/google-gemma-4-12b-local-ai/
-
https://winbuzzer.com/2026/06/02/microsoft-adds-seven-mai-models-to-foundry-for-developers-xcxwbn/
Microsoft is putting seven first-party MAI models into developer channels, led by the MAI-Thinking-1 reasoning model in Foundry private preview.
#AI #MAIThinking1 #MicrosoftFoundry #Microsoft #MicrosoftAI #AIModels #MultimodalAI #Build2026
-
https://winbuzzer.com/2026/06/02/microsoft-adds-seven-mai-models-to-foundry-for-developers-xcxwbn/
Microsoft is putting seven first-party MAI models into developer channels, led by the MAI-Thinking-1 reasoning model in Foundry private preview.
#AI #MAIThinking1 #MicrosoftFoundry #Microsoft #MicrosoftAI #AIModels #MultimodalAI #Build2026
-
https://winbuzzer.com/2026/06/01/nvidia-launches-cosmos-3-with-openmdw-for-physical-ai-xcxwbn/
NVIDIA has launched Cosmos 3 as a physical-AI model that combines scene reasoning, multimodal generation and action output, tying the release to a new OpenMDW licensing framework.
#AI #NVIDIA #PhysicalAI #AIModels #MultimodalAI #WorldModels #Robotics
-
https://winbuzzer.com/2026/06/01/nvidia-launches-cosmos-3-with-openmdw-for-physical-ai-xcxwbn/
NVIDIA has launched Cosmos 3 as a physical-AI model that combines scene reasoning, multimodal generation and action output, tying the release to a new OpenMDW licensing framework.
#AI #NVIDIA #PhysicalAI #AIModels #MultimodalAI #WorldModels #Robotics
-
https://winbuzzer.com/2026/06/01/minimax-launches-m3-with-1m-context-multimodal-push-xcxwbn/
MiniMax is pushing M3 into the long-context model race with multimodal input and a claimed 1 million-token window.
#AI #MiniMax #AIModels #MultimodalAI #AgenticAI #AICoding #MiniMaxM3 #ChinaAI
-
https://winbuzzer.com/2026/06/01/minimax-launches-m3-with-1m-context-multimodal-push-xcxwbn/
MiniMax is pushing M3 into the long-context model race with multimodal input and a claimed 1 million-token window.
#AI #MiniMax #AIModels #MultimodalAI #AgenticAI #AICoding #MiniMaxM3 #ChinaAI
-
https://winbuzzer.com/2026/05/25/bytedance-hkust-find-better-long-document-ai-training-xcxwbn/
ByteDance AI Training Study: Multimodal Q&A Strategy Beats Raw OCR-Transcription Input
#AI #ByteDance #HKUST #AIModels #AIResearch #AITraining #MultimodalAI
-
https://winbuzzer.com/2026/05/25/bytedance-hkust-find-better-long-document-ai-training-xcxwbn/
ByteDance AI Training Study: Multimodal Q&A Strategy Beats Raw OCR-Transcription Input
#AI #ByteDance #HKUST #AIModels #AIResearch #AITraining #MultimodalAI
-
Tornare da un viaggio significa quasi sempre ritrovarsi con una quantità ingestibile di foto. Nel caso di Lisbona, il problema non era tanto archiviare gli scatti, quanto riuscire a estrarne una ventina davvero condivisibile: belle, sì, ma anche varie e capaci di raccontare l’esperienza nel suo insieme. PhotoPrism offriva già un’ottima base grazie a geolocalizzazione, riconoscimento facciale, label e strumenti di organizzazione, ma non aveva ancora un modo per comporre automaticamente un album con “le foto più belle” e soprattutto con sufficiente varietà.
Da qui è nata l’idea di un selezionatore AI: una piccola applicazione Java che usa PhotoPrism per recuperare le miniature delle immagini e Ollama per far lavorare due modelli AI, uno multimodale per assegnare un punteggio estetico e produrre una descrizione oggettiva, e un secondo modello testuale per raggruppare semanticamente le foto e selezionarle con più equilibrio.
Il problema vero non era la qualità
Il primo prototipo faceva una cosa molto semplice: prendere le foto da PhotoPrism, inviarle a un modello multimodale su Ollama e chiedere un voto estetico da 1 a 100 insieme a una breve descrizione. Sulla carta sembrava sufficiente, ma in pratica produceva una selezione monotona: immagini molto belle singolarmente, ma spesso troppo simili tra loro.
Era il classico caso in cui un ranking puro ottimizza la qualità locale ma non la copertura narrativa. Se cinque foto dello stesso scorcio o dello stesso momento ricevono voti alti, un algoritmo ingenuo tende a sceglierle tutte. Per costruire un album da condividere, invece, non basta premiare le immagini migliori: bisogna anche evitare la ripetizione.
[…]
#ai #albumFotografici #clusteringSemantico #computerVision #fotografia #java #Maven #multimodalAI #ollama #organizzazioneFoto #photoprism #PhotoPrismAPI #selezioneFoto #selfHosted https://www.b0sh.net/2026/05/ho-costruito-un-selezionatore-ai-per-scegliere-le-foto-migliori-da-photoprism/ -
Tornare da un viaggio significa quasi sempre ritrovarsi con una quantità ingestibile di foto. Nel caso di Lisbona, il problema non era tanto archiviare gli scatti, quanto riuscire a estrarne una ventina davvero condivisibile: belle, sì, ma anche varie e capaci di raccontare l’esperienza nel suo insieme. PhotoPrism offriva già un’ottima base grazie a geolocalizzazione, riconoscimento facciale, label e strumenti di organizzazione, ma non aveva ancora un modo per comporre automaticamente un album con “le foto più belle” e soprattutto con sufficiente varietà.
Da qui è nata l’idea di un selezionatore AI: una piccola applicazione Java che usa PhotoPrism per recuperare le miniature delle immagini e Ollama per far lavorare due modelli AI, uno multimodale per assegnare un punteggio estetico e produrre una descrizione oggettiva, e un secondo modello testuale per raggruppare semanticamente le foto e selezionarle con più equilibrio.
Il problema vero non era la qualità
Il primo prototipo faceva una cosa molto semplice: prendere le foto da PhotoPrism, inviarle a un modello multimodale su Ollama e chiedere un voto estetico da 1 a 100 insieme a una breve descrizione. Sulla carta sembrava sufficiente, ma in pratica produceva una selezione monotona: immagini molto belle singolarmente, ma spesso troppo simili tra loro.
Era il classico caso in cui un ranking puro ottimizza la qualità locale ma non la copertura narrativa. Se cinque foto dello stesso scorcio o dello stesso momento ricevono voti alti, un algoritmo ingenuo tende a sceglierle tutte. Per costruire un album da condividere, invece, non basta premiare le immagini migliori: bisogna anche evitare la ripetizione.
[…]
#ai #albumFotografici #clusteringSemantico #computerVision #fotografia #java #Maven #multimodalAI #ollama #organizzazioneFoto #photoprism #PhotoPrismAPI #selezioneFoto #selfHosted https://www.b0sh.net/2026/05/ho-costruito-un-selezionatore-ai-per-scegliere-le-foto-migliori-da-photoprism/ -
https://winbuzzer.com/2026/05/22/gemini-adds-capcut-editing-as-google-expands-creation-xcxwbn/
Google is positioning Gemini as a bigger creative hub by bringing CapCut into the app alongside a broader partner push.
#AI #GoogleGemini #Gemini #Google #Capcut #GenAI #AITools #AIIntegration #AIAssistants #MultimodalAI #AIImageEditing
-
https://winbuzzer.com/2026/05/22/gemini-adds-capcut-editing-as-google-expands-creation-xcxwbn/
Google is positioning Gemini as a bigger creative hub by bringing CapCut into the app alongside a broader partner push.
#AI #GoogleGemini #Gemini #Google #Capcut #GenAI #AITools #AIIntegration #AIAssistants #MultimodalAI #AIImageEditing
-
https://winbuzzer.com/2026/05/20/gemini-app-rolling-out-neural-expressive-redesign-xcxwbn/
Google is rolling out a redesigned Gemini app while making Gemini 3.5 Flash the default model and integrating Gemini Spark and a Daily Brief feature.
#AI #GoogleGemini #GeminiApp #Google #GoogleAI #AIAssistants #AIAgents #AgenticAI #ConversationalAI #MultimodalAI
-
https://winbuzzer.com/2026/05/20/gemini-app-rolling-out-neural-expressive-redesign-xcxwbn/
Google is rolling out a redesigned Gemini app while making Gemini 3.5 Flash the default model and integrating Gemini Spark and a Daily Brief feature.
#AI #GoogleGemini #GeminiApp #Google #GoogleAI #AIAssistants #AIAgents #AgenticAI #ConversationalAI #MultimodalAI
-
https://winbuzzer.com/2026/05/20/google-adds-a-conversational-search-feature-to-you-xcxwbn/
YouTube is launching Ask YouTube as a conversational video search feature while adding Gemini Omni remixing into Shorts Remix and the YouTube Create app.
#AI #YouTube #Google #AskYouTube #GoogleGemini #GenAI #AISearch #MultimodalAI #GeminiOmni
-
https://winbuzzer.com/2026/05/20/google-adds-a-conversational-search-feature-to-you-xcxwbn/
YouTube is launching Ask YouTube as a conversational video search feature while adding Gemini Omni remixing into Shorts Remix and the YouTube Create app.
#AI #YouTube #Google #AskYouTube #GoogleGemini #GenAI #AISearch #MultimodalAI #GeminiOmni
-
https://winbuzzer.com/2026/05/20/google-launches-the-gemini-omni-multimodal-model-s-xcxwbn/
Google has launched Gemini Omni as a mixed-input AI model family and started the first live rollout through Gemini Omni Flash on Gemini, YouTube, and creator surfaces.
#AI #GoogleGemini #Google #GeminiOmni #GoogleAI #GoogleDeepMind #MultimodalAI #AIModels #AIVideoGeneration #AIVideo #TextToVideo #GenAI
-
https://winbuzzer.com/2026/05/20/google-launches-the-gemini-omni-multimodal-model-s-xcxwbn/
Google has launched Gemini Omni as a mixed-input AI model family and started the first live rollout through Gemini Omni Flash on Gemini, YouTube, and creator surfaces.
#AI #GoogleGemini #Google #GeminiOmni #GoogleAI #GoogleDeepMind #MultimodalAI #AIModels #AIVideoGeneration #AIVideo #TextToVideo #GenAI
-
https://winbuzzer.com/2026/05/18/oppo-open-sources-x-omniclaw-on-device-android-ai-xcxwbn/
Oppo has published the X-OmniClaw as an open-source Android AI agent that works from the phone itself.
#AI #XOmniClaw #Oppo #AndroidAI #Android #AIAgents #OnDeviceAI #MultimodalAI
-
https://winbuzzer.com/2026/05/18/oppo-open-sources-x-omniclaw-on-device-android-ai-xcxwbn/
Oppo has published the X-OmniClaw as an open-source Android AI agent that works from the phone itself.
#AI #XOmniClaw #Oppo #AndroidAI #Android #AIAgents #OnDeviceAI #MultimodalAI
-
THE WHIRRING MACHINERY OF WORDS: NAVIGATING THE LLM LANDSCAPE
New large language models can now understand and create text, images, and audio. This helps researchers and businesses.
-
THE WHIRRING MACHINERY OF WORDS: NAVIGATING THE LLM LANDSCAPE
New large language models can now understand and create text, images, and audio. This helps researchers and businesses.
-
New AI models can now work with text, images, and sound, unlike older models that only used text. This is a big step forward.
#AI, #LLM, #TechNews, #Innovation, #MultimodalAI
https://newsletter.tf/llms-understand-text-images-audio/ -
New AI models can now work with text, images, and sound, unlike older models that only used text. This is a big step forward.
#AI, #LLM, #TechNews, #Innovation, #MultimodalAI
https://newsletter.tf/llms-understand-text-images-audio/ -
https://winbuzzer.com/2026/05/13/thinking-machines-wants-to-build-an-ai-that-actual-xcxwbn/
Thinking Machines Lab has previewed a research-stage full-duplex AI system built to keep listening while it responds, rather than waiting for turn-based exchanges.
#AI #ThinkingMachinesLab #MiraMurati #VoiceAI #AIModels #ConversationalAI #MultimodalAI #VoiceAssistants
-
https://winbuzzer.com/2026/05/13/thinking-machines-wants-to-build-an-ai-that-actual-xcxwbn/
Thinking Machines Lab has previewed a research-stage full-duplex AI system built to keep listening while it responds, rather than waiting for turn-based exchanges.
#AI #ThinkingMachinesLab #MiraMurati #VoiceAI #AIModels #ConversationalAI #MultimodalAI #VoiceAssistants
-
NVIDIA Nemotron 3 Nano Omni: Open Multimodal AI Agent Guide 2026
NVIDIA released Nemotron 3 Nano Omni on April 28, 2026 — the first open model to natively unify vision, audio, and language in a shared reasoning loop, delivering 9x highe...
https://wowhow.cloud/blogs/nvidia-nemotron-3-nano-omni-multimodal-agent-developer-guide-2026
-
https://winbuzzer.com/2026/05/11/gemini-api-file-search-is-now-multimodal-xcxwbn/
Google has expanded Gemini API File Search with multimodal retrieval, metadata filtering, and page citations.
#AI #GeminiAPI #Google #GoogleGemini #GoogleAI #AITools #AISearch #MultimodalAI