Chris @chickenmcdev
๐ฆ๐ท From the land of asado. Home lab dev | Small LLMs on 16GB VRAM | Agentic coding & tech enthusiast Argentina Joined September 2026-
Tweets34
-
Followers0
-
Following148
-
Likes153
@atomic_chat_hq That was fast ๐โโ๏ธ๐จ
Introducing ๐ป๐ต๐ผ๐น: our new generation of Liquid Foundation Models. With a new tri-hybrid architecture and new multimodal encoders, the LFM3 family is the ideal model suite for specialized intelligence. > 6 model instances, across text, audio, vision, and omni modalities, generative and decision models. > ๐ป๐ต๐ผ๐น-๐ท.๐ฝ๐ฑ: text, 256k context window. > ๐ป๐ต๐ผ๐น-๐ ๐ป-๐ธ๐ฑ: vision-language model with video-input support for spatial understanding > ๐ป๐ต๐ผ๐น-๐ฐ๐๐๐๐-๐ธ๐ฑ: tool calls, voice chat, reasoning in 18 languages. > ๐ป๐ต๐ผ๐น-๐พ๐๐๐-๐ธ๐ฑ: an omni model combines vision, audio, and language. > ๐๐ท-๐ธ๐ฑ and ๐๐ท-๐พ๐๐๐-๐ธ๐ฑ: decision models with typed outputs and calibrated probabilities in a single pass. Omni adds audio input. > fine-tune on 1 GPU. > a new tri-hybrid architecture > architecture optimized for 14 processor classes across GPUs, CPUs, and NPUs. > ๐ป๐ต๐ผ๐น-๐ ๐ป-๐ธ๐ฑ is a beast. It performs competitively with Gemma 4 12B (6x larger), and Gemini 3.5 Flash Lite, and GPT-5.6 Luna, with 1.9B parameters. > ๐ป๐ต๐ผ๐น-๐ท.๐ฝ๐ฑ is #1 on the Pareto frontier of device intelligence ๐๐๐๐๐๐๐.๐๐๐๐๐๐.๐๐ Every modality for specialized intelligence. Pick the one that fits your use case, customize it, and run it on-device. 1/ ๐งต
๐จ Confirmed. OpenCode hacked. Keys stolen. Telling customers to revoke.
Seems OpenCode customers getting notices of a breach. Got some DMs that peopleโs API keys are running up charges. Not seeing it on the news yet. Anyone else?
Thanks for your hard work !
Today, weโre releasing Bonsai 2 2B, 4B, and 9B. Based on Qwen3.5, these new ternary models bring stronger reasoning, visual understanding, and tool calling to compact footprints designed for everyday devices: โข 2B: 0.43GB โข 4B: 0.95GB โข 9B: 2.0GB For this release, we
DeepSeek V4.1 Flash is now free in OpenCode
Today we release ๐ป๐ต๐ผ๐ธ.๐ป-๐ธ๐บ๐ฑ-๐ฐ๐ธ๐ฑ, the largest ๐ป๐ต๐ผ๐ธ.๐ป family member. It is an efficient local model for laptops and PCs to work through multi-step tasks inside popular agent harnesses. > 24B total parameters, 2B active per token > Context expanded to 131K tokens > A 128K-token vocabulary > 33T pretraining tokens > It is a reasoning model > Very easy to fine-tune and customize > Performs competitively with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks. > Open base and post-trained weights on @huggingface ๐ค 1/
๐ Meet Qwen-Image-2.1-Turbo โ create and edit images in just 8 denoising steps! Open weights now available! Built on Qwen-Image-2.1, Turbo is an accelerated checkpoint on the same 7B visual generation architecture. Fewer steps does not mean lower quality: it still generates strong 2K images from text, and supports continued creation through natural-language edits, from adding accessories to changing a scene. Start directly with Diffusers: load QwenImage21Pipeline and the checkpointโs recommended 8-step sampling schedule is ready to go. ๐ Prefer an API? Qwen-Image-2.1 Pro and Turbo APIs are now officially live. Choose a hosted API for your application, or use the Turbo weights in your own workflow. ๐ค ModelScope: modelscope.cn/models/Qwen/Qwโฆ ๐ค Hugging Face: huggingface.co/Qwen/Qwen-Imagโฆ ๐ Pro API: modelstudio.console.alibabacloud.com/ap-southeast-1โฆ ๐ Turbo API: modelstudio.console.alibabacloud.com/ap-southeast-1โฆ What will you create with 8 steps? Share your results with us!
Meet Qwen-Image-2.1, the most balanced and cost-effective image generation model in the Qwen-Image series! Now open weights! ๐จ A unified model for both generation and editing, delivering top-tier quality in a lightweight package. Highlights: ๐ - Compact & exceptionally fast:
Introducing Claude Haiku 5.5: the cheapest, fastest, and most capable small model weโve ever released. On average, it costs around 75% less to run than Claude Haiku 4.5.
Introducing EmbeddingGemma 2, a new open multimodal model that sets the standard for on-device efficiency. - our first open, natively multimodal embedding model - handles text, code, image, video, and audio tasks within a lightweight, modular 740M parameter form factor - ideal for offline, privacy-first RAG when paired with Gemma 4 - outperforms some specialist models more than twice its size Weights available now on Hugging Face.
Another project in V: HorneroOS, an Arch-based Hyprland desktop for the Agentic Era, focused on deep customization, themes, layouts and polished UX. Its core tooling is written in V. Built by @ulisesjcf horneroos.com github.com/HorneroOS/hornโฆ
@RoundtableSpace Reset? reset token usage you mean?
@daveopssh @pidotdev @opencode Instalate herdr y a combinar los dos ๐
@maxifirtman pensar que hoy la gente usa chatgpt como terapia ya ๐
๐ 120 t/s using coding config: Ornith-1.5-9B-Q4 + Ornith-1.5-9B-DFlash-Q4_K_M RTX 4060ti 16gb 32GB DDR4 -c 131072 --spec-type draft-dflash --spec-draft-n-max 6 --temp 0.6 --top-p 0.95 --top-k 20 --min-p 0.0 --repeat-penalty 1.0 --presence-penalty 0.0 ๐ค from @ornith_
๐ชฝDFlash draft models are available for Ornith-1.5 9B, 35B-A3B, and 397B. DFlash and Ornith 1.5 make a dynamic duo: DFlash performs as the draft model, accelerating inference by drafting multiple tokens at once. Then Ornith 1.5 verifies the draft tokens in one pass instead of
@maxifirtman al final los modelos van a terminar como los misiles. Gana el que tenga la mayor y mejor cantidad pero con la promesa de no usarlo contra el otro.
GPT-6.1 Sol: near-Astra intelligence for a fifth of the price. Itโs the most cost-efficient model for its performance available today.
same jev example from @typesafeai during @OpenAI decisions api presentation?
Angeles @aquela_cba
3K Followers 2K Following Ethical Hacker | AI & Software Dev | AI Security Research | Building aquela ๐บ adversalia ๐ผ & rootwomen https://t.co/gCIAR9BJg0 https://t.co/K8Qk00cOfa
Chris ๐จ๐ฆ @llm_wizard
5K Followers 784 Following Open Source Model Lover @ NVIDIA AI Views my own.
Calif @calif_io
23K Followers 34 Following We're https://t.co/KTEDnC3tKt. Join us to make the Internet safer for your mum and everyone else: https://t.co/eUFMLkWHiA.
The PHP Foundation @ThePHPF
27K Followers 0 Following We support, promote, and advance the PHP language. ๐๐ #ThePHPFoundation Donate at https://t.co/B0te9FaPEa
SGLang @sgl_project
12K Followers 55 Following Run LLMs fast at any scale ๐ https://t.co/F3u6wYESL0 Join our community https://t.co/fmlOfTOEec For AI tech blogs & deep-dives ๐ @lmsysorg
Zed @zeddotdev
86K Followers 109 Following We are Zed Industries. - Delta: A multiplayer environment for coding with agents, powered by DeltaDB. - Zed: The world's fastest IDE. Open source. AI optional.
fal @fal
70K Followers 0 Following the generative media cloud. hiring https://t.co/JrbUk989MN. for support/discounts, e-mail us at [email protected].
OpenBMB @OpenBMB
13K Followers 397 Following OpenBMB (Open Lab for Big Model Base) aims to build foundation models and systems towards AGI. Connect with us: https://t.co/N9pevTnoOa
OpenDesign @OpenDesignHQ
26K Followers 66 Following The vibe design workspace for everyone and every team. 90K+ GitHub โญ ยท open-source ยท by any agent
GMI Cloud @gmi_cloud
10K Followers 62 Following one cloud for compute, inference, and agents Question: https://t.co/dOcscX4hmx Japan: @gmicloudJapan
OpenBot @OpenBot_
13K Followers 10 Following Workspace for AI coworkers. Open-source & free to start. Use it with any AI models you already have.
Armin Ronacher โ @mitsuhiko
99K Followers 912 Following Creator of Flask. Building at https://t.co/uGuzfu0LKT. Bypassing Permissions. Can hand crank. Husband and father of 3 โ โmore nuanced in personโ
OpenRouter @OpenRouter
156K Followers 389 Following The unified interface for LLMs. 400+ models (incl. 50+ free), the best prices, explorable data, private chat, & a unified API. https://t.co/qJG5mKrQ6j
Hugging Face @huggingface
831K Followers 227 Following The AI community building the future. https://t.co/TpiXQMQ9rZ
Orchestrator.inc @useOrchestrator
3K Followers 18 Following Manage all your coding agents effortlessly โญ 13K+ on https://t.co/oFCXVJPE3d ๐ง 2.3K+ on https://t.co/DQTiVNqCVj
The V Programming Lan... @v_language
12K Followers 7 Following Simple, fast, safe, compiled language for developing maintainable software.
Mia @MiaAI_lab
44K Followers 382 Following Building with AI & LLMs // Running frontier open models locally
T3 Code @t3dotcodes
11K Followers 2 Following A better agentic code app. Built by @theo and @jullerino
Kubernetes Community ... @kcdargentina
126 Followers 13 Following Comunidad Argentina de cloud native foundation
Neon @neondatabase
21K Followers 822 Following Backend for apps and agents built on Lakebase Postgres Free Plan: https://t.co/2oFkBiuLiV
Sergio Paniego @SergioPaniego
7K Followers 2K Following Machine Learning Engineer @huggingface ๐ค AI PhD. Technology enables us to be more human. ๐ณ๏ธโ๐
TinyHumans AI @tinyhumansai
8K Followers 95 Following An AI lab focused on building AI agents with human-like AI memory and an artificial subconscious, community: https://t.co/yW743nW2MN.
Termius @TermiusHQ
13K Followers 1 Following Transform your workflow with Termius โ Modern SSH client designed for team collaboration, productivity, and a seamless experience across devices.
Pirate Face @thepirateface
13K Followers 83 Following Secure the future of open weight models. Turn AI into torrents that live forever.
Ornith @ornith_
20K Followers 391 Following Once you have tasted flight, you will walk the earth with your eyes turned skywards.
TheStage AI @TheStageAI
451 Followers 340 Following Automated Enterprise Inference Stack & Research Lab
Aseprite @aseprite
96K Followers 88 Following I'm a little sprite editor ๐ @igarastudio feeds me๐ฅ๐ช๐ฎ #aseprite
Pi Changelog @PiChangelog
6K Followers 4 Following Unofficial Pi release updates creds @badlogicgames @pidotdev
James Grugett @jahooma
13K Followers 587 Following Founder of Freebuff & Manifold Markets Let's allow more experiments, create more positive-sum exchanges, and set more ambitious goals.
Arena.ai @arena
238K Followers 224 Following Where AI meets the real world. We measure and advance the frontier of AI through community-driven evaluation. Weโre hiring โ https://t.co/XBZCrsdD77
Aye ๐ฉ๐ปโ๐ป ๏ฟฝ... @ayebytes
6K Followers 882 Following explico bitcoin desde el cรณdigo y comparto mis experimentos con ai. ingeniera de sistemas. hice @satlensapp ๐ฆ๐ท
ใใซใป ใขใค(Ai-H... @ai_hakase_
19K Followers 815 Following ็ๆAIใจ็ซใธใฎๆใงๅใใฆใใAI้็บไผ็คพใฎIPใงใ๐พ ComfyUIใฎ้ซๅบฆใชใฏใผใฏใใญใผๆง็ฏใๆๆฐใขใใซๆค่จผใๅฎๅ็ด็ตใฎAIใจใผใธใงใณใๆดป็จๆณใ็บไฟกไธญใ ๆ้ ใฎๅ จๆใจใใกใคใซใฏ Note ใธ๐ ไปใซใ YouTube๏ฝPatreon ใใใฃใฆใพใ ๐ ใๅฝน็ซใกๆ ๅ ฑใฏใใใคใฉใคใใใซ้็ดไธญ๐ฅณ
SparkLLM @SparkLLM
2K Followers 26 Following Leading general-purpose foundation models & professional AI agents. The proprietary Spark model family, with AStudio & Loomy harnesses. Open source.
Orca @orca_build
22K Followers 195 Following Run a team of agents any subscription Open source (MIT) โข Desktop + mobile + VPS Repo: https://t.co/NW8tpPUBw4 (108k โญ๏ธ) Discord: https://t.co/BOypIwh6Li
Jinjing Liang @JinjingLiang
11K Followers 2K Following building @orca_build (105k โญ) prev @google @cornell, YC alum
IQuest @IQuest_research
154 Followers 22 Following Building AI that codes the future. From compute systems to intelligent agents โ powering the next leap in life sciences.
Trends for United States
You might like
