Top Open-Source AI Models launches from 2026, ranked by X views.
Over the past two months we open-sourced a set of on-device audio models. The name is Audio8. The series can run on phones, PCs, and other hardware with limited compute and memory. You can use them for local inference right away. What’s included: ASR: 0.1B, 0.3B, 0.6B, 3B TTS: https://t.co/eWnEWmoA4H
Seedance 2.0 is now available to ALL users. Create studio-quality videos by directing text, images, video, and audio into a single unified scene with perfectly synchronized stereo sound and realistic physics. Hollywood-level production in less than 2 minutes. https://t.co/2ojXadAft0
Nano Banana 2 Lite is now available in Runway. Create images at warp speed, without compromising on quality. Get started at the link below or ask Agent to use Nano Banana 2 Lite. https://t.co/FlugLkeNmz
Today, Baseten and @MicrosoftAI are excited to announce that MAI-Thinking-1 is coming to Baseten. MAI-Thinking-1 is a model you can fine-tune without giving your data to the lab. Key characteristics include: → Clean data lineage, with zero distillation from third-party models https://t.co/ulqekRVT17
Need document parsing that stays fully local and private? 👀 Meet liteparse-server, a self-hostable, open-source HTTP server for parsing documents and generating screenshots from PDFs, Office files, and images. ✅ 100% self-hosted ✅ Private by default ✅ Open source ✅ Built https://t.co/nYe2VYroBX
introducing Pinterest in Krea Agent. now you can bring all your pins and boards into the tool and use them as context for your agents. try it now at krea . ai / agent https://t.co/dy6tGvRwBE
Open source 🤝 NVIDIA https://t.co/9JqtS7FNhD
The 4-bit quant is now available, making our first open-source agentic coding model small enough to be run on a Mac. Get the weights here: https://t.co/8amCVzzunE
Introducing Astra 2. this high-quality video enhancer comes with fine-grained controls like creativity, realism, or sharpness. try it now👇 https://t.co/m5iaMnbc4C
It's here! Introducing MiniMax H3-Image, and we're releasing the model weights! 🔥16 million pixel. 10-20 seconds. Open weights. Built on H3, our new image model brings together three things that matter most: ⚡ Speed: Generate stunning 4K × 4K images in 10-20 seconds. 🎨 https://t.co/M43849t54s
introducing Ideogram v4.0. 2k native resolution, excellent text rendering, and support for JSON prompts. try it now in Krea. https://t.co/0U2ypOFHlW
Let's talk parsing charts 📊📈. Last week we released ParseBench, the first document OCR benchmark for AI agents. New in ParseBench: ChartDataPointMatch. Most document look at a chart and OCR the caption. Agents need the actual numbers. That's the gap between "OCR'd the https://t.co/xlXrANpHhF
Today we're introducing Claude for Founding. Just enter your prompt and Shipper oneshots a full business making you money in one click. This is truly insane. https://t.co/nU26icPGcS
Check out everything we launched today 👇 https://t.co/Mck7gcPr4c
MiniMax-M3 from @MiniMax_AI is now available on Together AI. It’s an open-weight native multimodal model with 1M context, MiniMax Sparse Attention, and thinking / non-thinking modes. Together AI is MiniMax’s preferred cloud partner, with inference optimizations delivering up to https://t.co/kX7UvXLeEW
Krea 2 LoRAs are now available to everyone. try this one below 👇 https://t.co/JiZEil1bqx
We tested @claudeai Opus 4.8 (High) on APEX-SWE ahead of today's release. It's the new #1 at 45.3% Pass@1, nearly 4 points ahead of GPT-5.3 Codex (41.5%). Congrats @AnthropicAI on the release and having three models in the top 5! https://t.co/oIcn5wNvTN
We’ve fully revamped our LlamaCloud integration with @n8n_io as part of the LlamaParse v2 + new LlamaCloud SDK launch. What’s new: ✅ Fully stable nodes for LlamaParse, LlamaExtract, and LlamaCloud Index 🆕 Two new services: LlamaClassify and LlamaSheets 📘 A complete https://t.co/4L8y1F74uw
Today, we're introducing Claude for business building. Just enter your website and it will code, design, launch and translate a mobile app inspired by the original website. This is just insane. https://t.co/GubnNgetLv
May means more models. Today we're introducing Anthropic's Claude Opus 4.8, available now in Copilot Cowork (Frontier) and rolling out to Copilot Chat, Excel, PowerPoint, and Copilot Studio early release cycle environments.
We just released the DSpark weights for Nanbeige4.2-3B (https://t.co/8lWoxcgYqI). https://t.co/lXqTqkCcez
The incredible 4.0 image models from @ideogram_ai are here. Native 2K resolution + strong typography, and completely open source. https://t.co/9ZlfZwR5Hm https://t.co/u1NF4IxD0s
Gemini 3.5 Flash is now available on Replicate via API! https://t.co/K54shrwkED https://t.co/XCWzqkYHbE https://t.co/qb8OAFUHiU
introducing moodboard sharing. you can now share Krea 2 moodboards with others. a few you can try below 👇 https://t.co/lpPW1Q9anv
Introducing GLM-5.2 from @Zai_org, https://t.co/Pp24eiUBaD’s latest flagship open model for long-horizon tasks with 1M context, flexible thinking effort, and stronger agentic coding. Now available on Together AI, GLM-5.2 runs on research-powered inference for long-context, https://t.co/3BKhEwrysY
ExtractBench is now live on @Kaggle. It tests schema-guided document extraction on the documents most likely to break downstream agents and workflows, including long record lists, noisy scans, handwriting, and complex tables. The benchmark covers 370 enterprise documents across https://t.co/rTDYMUABll