Image created with gemini-3.1-flash-image-preview with claude-sonnet-4-5. Image prompt: Wide-angle Jia Zhangke style documentary still of abandoned border checkpoint at dawn, muted gray-blue palette, half-demolished concrete booths with peeling paint and rusted barriers, dirt road fading into humid haze, one calm chestnut horse standing naturally between structures in middle distance, lone guard in worn civilian jacket watching from doorway, overcast flat natural light, desaturated colors, observational realism, large white text overlay reading INTERNATIONAL in Chinese cinema poster style, 35mm digital texture, decelerated stillness, transitional atmosphere
Exclusive | Pentagon Used Anthropic’s Claude in Maduro Venezuela Raid – WSJ https://www.wsj.com/politics/national-security/pentagon-used-anthropics-claude-in-maduro-venezuela-raid-583aff17
Anthropic is prepared to loosen its current terms of use, but wants to ensure its tools aren’t used to spy on Americans en masse, or to develop weapons that fire with no human involvement. The Pentagon has aid, that Anthropic will “”pay a price”” for that behavior. Within this”” https://x.com/kimmonismus/status/2023419652378955809
You can now run Qwen3.5 locally! 💜 Qwen3.5-397B-A17B is an open MoE vision reasoning LLM for agentic coding & chat. It performs on par with Gemini 3 Pro, Claude Opus 4.5 & GPT-5.2. Run 4-bit on 256GB Mac / RAM. Guide: https://t.co/wjS1lMnbNp GGUF: https://x.com/UnslothAI/status/2023338222601064463
BitDance is our lunar new year gift: a 14B parameters autoregressive image generation model – that is autoregressive in bits, not codebooks It’s fast given 14B parameters, try it: https://x.com/multimodalart/status/2023797260057014372
ByteDance releases BitDance: Scaling Autoregressive Generative Models with Binary Tokens “”We present BitDance, a scalable autoregressive (AR) image generator that predicts binary visual tokens instead of codebook indices. With high-entropy binary latents, BitDance lets each”” https://x.com/iScienceLuvr/status/2023707945104458097
🎉 Congrats to @Alibaba_Qwen on releasing Qwen3.5 on Chinese New Year’s Eve — day-0 support is ready in vLLM! Qwen3.5 is a multimodal MoE with Gated Delta Networks architecture — 397B total params, only 17B active. What makes it interesting for inference: 🧠 Gated Delta”” https://x.com/vllm_project/status/2023341059343061138
🔥 Alibaba’s Qwen 3.5 just dropped — and Zhihu is dissecting it. Here’s a sharp breakdown from Zhihu contributor toyama nao 👇 🏆 Verdict: “”The spearhead of the open-source elite.”” 📊 Big picture Tongyi Lab’s pattern: new mid-size model leapfrogs old giant. • Last cycle: 80B”” https://x.com/ZhihuFrontier/status/2024176484232155236
Qwen https://qwen.ai/blog?id=qwen3.5
Qwen3.5 is Live! Today we openweight the first model, Qwen3-397B-A17B, which is a native multimodal model supporting both thinking and non-thinking modes. We have strengthened its coding and agentic capabilities to foster productivity for developers and enterprises. Hope you”” https://x.com/JustinLin610/status/2023332446713070039
Alibaba Yunqi: 7 models released in 4 days (Qwen3-Max, Qwen3-Omni, Qwen3-VL) and $52B roadmap | AINews https://news.smol.ai/issues/25-09-23-alibaba-yunqi
Alibaba’s new Qwen3.5-397B-A17B is the #3 open weights model in the Artificial Analysis Intelligence Index – a significant upgrade from Qwen3-235B-A22B-2507, and achieved with fewer active parameters than leading peers Qwen3.5-397B-A17B is the first model released by Alibaba”” https://x.com/ArtificialAnlys/status/2023794497055060262
Qwen https://qwen.ai/blog?id=qwen3.5#spatial-intelligence
Qwen3.5’s thinking is downright excessive.”” https://x.com/QuixiAI/status/2023995215690781143
A short film Jia Zhangke produced using Seedance 2.0 for Chinese New Year (Subtitled) – YouTube
As promised, here’s the short film Jia Zhangke produced using Seedance 2.0 for Chinese New Year and his take on AI filmmaking”” https://x.com/FrankYan2/status/2023257752017981446
ByteDance Seed https://seed.bytedance.com/en/seed2
Jia Zhangke is considered one of the greatest living directors… not an AI hack. He’s worth Googling. I’m overwhelmed by the critical acclaim he has, and his brazen “look at me make an AI video” flex. The X video has subtitles, and the YouTube one has commentary in the description from Vanity Fair.
Primer: Where to start with Jia Zhangke
https://www.avclub.com/primer-jia-zhangke
Ranking the Jia Zhangke Films – The Reel World
the first official AI movie is here and.. its wild China’s top director Jia Zhangke was so impressed by Seedance 2.0 that he made a film himself.. in just 3 days when asked if AI will replace filmmakers, he said cinema has always moved with tech. Digital cameras didn’t kill”” https://x.com/EHuanglu/status/2023449238114320514
Where to begin with Jia Zhangke | BFI
https://www.bfi.org.uk/features/where-begin-jia-zhangke
WORLD APART: THE FILMS OF JIA ZHANGKE
https://www.artforum.com/features/world-apart-the-films-of-jia-zhangke-171661/
Koyeb is Joining Mistral AI to Build the Future of AI Infrastructure – Koyeb https://www.koyeb.com/blog/koyeb-is-joining-mistral-ai-to-build-the-future-of-ai-infrastructure#serverless-inference-and-agents
BREAKING 🚨: Cline released Cline CLI 2.0, an open-source AI coding agent powered by Kimi K2.5 and MiniMax M2.5, available for free! `npm install -g cline` on Windows, Mac, and Linux. Open-source strikes back 👀”” https://x.com/testingcatalog/status/2022348951459172604
Introducing Cline CLI 2.0: An open-source AI coding agent that runs entirely in your terminal. Parallel agents, headless CI/CD pipelines, ACP support for any editor, and a completely redesigned developer experience. Minimax M2.5 and Kimi K2.5 are free to use for a limited time.”” https://x.com/cline/status/2022341254965772367
DeepSeek-V3.2 on GB300: Performance Breakthrough | vLLM Blog https://vllm.ai/blog/gb300-deepseek
People who say “”ohh this model is faster it must be mini”” must have forgotten that DeepSeek boasted of 60 tps when V3 came out. V3, of course, was ≈3x larger and had almost 2x more active params than V2. Then they increased the batch size and tps went back down. it’s weak signal”” https://x.com/teortaxesTex/status/2022255213394948360
Tiny Aya is out on Hugging Face a family of massively multilingual small language models Tiny Aya delivers strong multilingual performance in 70+ global languages in a 3.35B parameter model”” https://x.com/_akhaliq/status/2023771434347044890
Moondream’s inference engine got so fast image decoding became a bottleneck. So we shipped a SIMD image decoding library that’s faster than all the Python options I know of. Plus it’s statically linked so not a pain in the ass to install.”” https://x.com/vikhyatk/status/2024005498874306984
All the important news from the ongoing India AI Impact Summit | TechCrunch https://techcrunch.com/2026/02/22/all-the-important-news-from-the-ongoing-india-ai-summit/
Chinese New Year is rapidly becoming the AI researcher’s favorite holiday”” https://x.com/iScienceLuvr/status/2023312965756449088
Impressive benchmarks for the new Chinese LLM. The system card notes some gaps with US closed source models in code generation & wide knowledge, so be interested to see it in operation. Not clear it is open weights though? If not, represents a large shift in the AI market.”” https://x.com/emollick/status/2022658647378268361
Europe is funding new AI labs. No equity. Up to €125M. Munich is the first room. On March 4 the Next Frontier AI roadshow starts in Munich. The idea is simple: gather people who actually want to build new foundations for AI and put them together early enough that real teams”” https://x.com/IlirAliu_/status/2022352128438624427
Ireland opens probe into Musk’s Grok AI over sexualised images | Reuters https://www.reuters.com/sustainability/boards-policy-regulation/ireland-opens-probe-into-musks-grok-ai-over-sexualised-images-2026-02-17/
Chinese are now making AI edits of MacArthur as vengeful horror movie antagonist, taking out his wrath on Emperor Hirohito”” https://x.com/escapefrommelos/status/2023085808157175917
Paper : https://t.co/WvJL8tJ2IL Model weights : https://t.co/IQ3b0CjEeF Mistral AI studio live playground :”” https://x.com/GuillaumeLample/status/2024445952812060715
Mistral AI buys Koyeb in first acquisition to back its cloud ambitions | TechCrunch https://techcrunch.com/2026/02/17/mistral-ai-buys-koyeb-in-first-acquisition-to-back-its-cloud-ambitions/
Nice standalone Swift package for real-time streaming transcription with Mistral’s Voxtral Mini 4B in MLX Swift:”” https://x.com/awnihannun/status/2022322714548338962
also have fast lanczos3 resize but it doesn’t beat pyvips yet”” https://x.com/vikhyatk/status/2024008173271863541
I made Moondream’s MoE inference kernel 2.6% faster, without touching any kernel code!”” https://x.com/vikhyatk/status/2023749843186078144
1/ People often think better multilingual models must come at the cost of English performance. Not true. The constraint isn’t capacity, it’s data quality, and we can fix it. Today @datologyAI shares ÜberWeb: a year of multilingual curation lessons, scaled to 20T+ tokens.”” https://x.com/RicardoMonti9/status/2024136992779559055
This is false If no sanctions were present, then the gap would either be super small or non-existent DeepSeek bros (@zheanxu & @chenggang_zhao) would absolutely cook on Rubin & Blackwells”” https://x.com/zephyr_z9/status/2024437158988353630
This matches the general feeling on the big Chinese open source models. They have great benchmarks and near-frontier status on some coding, but there is a larger gap with the the big closed models than the benchmarks would indicate when it comes to real work and general “smarts””” https://x.com/emollick/status/2024190674166239420
Oof, SWE-rebench is brutal for recent Chinese releases M2.5 reported 80.2% on SWE-bench verified against 80.8% for Opus 4.6, but it seriously underperforms here Qwen3-Coder-Next looks good with 40% and only 80B A3B parameters”” https://x.com/maximelabonne/status/2022401174549512576
🚀 Qwen Coding Plan is now live on Alibaba Cloud Model Studio! ✨ What you get: • 🔥 Latest Qwen3.5-Plus models • 💡 Fixed monthly subscription: from ~$10/mo (Lite) or ~$50/mo (Pro) • 📦 Up to 90K requests/month for AI-powered coding • 🔌 Works with Claude Code, Qwen Code,”” https://x.com/Alibaba_Qwen/status/2024136381308805564
Ouch, the pricing on Alibaba just hurts. You can get the larger Kimi-K2.5 and GLM-5 for less”” https://x.com/scaling01/status/2023346718377406840
Qwen3.5-397B-A17B SVG results I have seen better. DeepSeek-V3.2 and GLM-5 both beat it.”” https://x.com/scaling01/status/2023364296277721300
🚀 Qwen3.5-397B-A17B is here: The first open-weight model in the Qwen3.5 series. 🖼️Native multimodal. Trained for real-world agents. ✨Powered by hybrid linear attention + sparse MoE and large-scale RL environment scaling. ⚡8.6x-19.0x decoding throughput vs Qwen3-Max 🌍201″” https://x.com/Alibaba_Qwen/status/2023331062433153103
Happy Chinese New Year!! What a week for open-source LLMs: > Qwen-3.5 > GLM-5 > MiniMax-M2.5 Are we just waiting on DeepSeek-V4 now? Also I’m hoping a US lab steps up with a true frontier open-source model.”” https://x.com/Yuchenj_UW/status/2023453819938763092
Qwen 3.5 goes bankrupt on Vending-Bench 2″” https://x.com/andonlabs/status/2023450768406364238
So a new Repo full of MLX-LM-LoRA examples to train your own LLM for Apple Silicon, fast and efficient on ultra long context lengths: Fine-tune Qwen3 4B Instruct on 32K context: https://t.co/yGZlR59fHD Train @IBMResearch Granite 350M model on RL-GRPO Reasoning:”” https://x.com/ActuallyIsaak/status/2022414004623479014
🚀 Qwen3.5-397B-A17B-FP8 weights are now open! It took some time to adapt the inference frameworks, but here we are: ✅ SGLang support is merged 🔄 vLLM PR submitted → https://t.co/rJkuitOBWs Check the model card for example code. vLLM support landing in the next couple of days!”” https://x.com/Alibaba_Qwen/status/2024161147537232110
🚩Cerebras’s MiniMax-M2 GGUF 2-bit model: https://t.co/udlviJQZqQ Qwen3-Coder-Next INT4 model:”” https://x.com/HaihaoShen/status/2022293472796180676
A clarification of Qwen3.5 Plus and 397B: 1. for opensource, we follow the tradition to make parameters apparent so we use the name with the number of total parameters and active params. 2. Qwen3-Plus is a hosted API version of 397B. As the model natively supports 256K tokens,”” https://x.com/JustinLin610/status/2023340126479569140
It’s Qwen 3.5 day today! 🥳 State of the art 800 GB model. Runs _locally_ with MLX using Q4, taking 225 GB of RAM.”” https://x.com/pcuenq/status/2023369902011121869
Let’s do the KV cache math for Qwen3.5: – KV heads: 2 – Head dimension: 256 – gated attention layers: 15 – bytes per element (BF16): 2 2 x 256 x 15 x 2 = 15 360 This is the same for K and V. So, we multiply by 2: 30 720 bytes Roughly 31 kb per token of context. Meaning at max”” https://x.com/bnjmn_marie/status/2023424404504342608
ollama run qwen3.5:cloud Qwen3.5-397B-A17B is the first open-weight model in the series. It’s available on Ollama’s cloud right now! Give it a try. Let’s go! 🚀🚀🚀”” https://x.com/ollama/status/2023334181804069099
Qwen 3.5 Plus is now available on AI Gateway. Thanks @vercel_dev team. 🤝 Use model: ‘alibaba/qwen3.5-plus’ Try it now!”” https://x.com/Alibaba_Qwen/status/2024029499541909920
Qwen3.5 runs quite well in mlx-lm. Awesome that we have a frontier-level hybrid model. The context gets longer but the inference speed and memory use barely change. Here’s the Q4 generating a space invaders game on an M3 Ultra. Generated 4,120 tokens at 37.6 tok/s.”” https://x.com/awnihannun/status/2023462412092059679
So speaking of benchmarks, what can be said of the new open Qwen? First, it completely destroys Qwen3-VL-235B ofc, but more surprisingly it outscores Qwen3-Max-thinking. All the while it’s the same model as “”Plus””. Plus just has 1M context and some more bells and whistles.”” https://x.com/teortaxesTex/status/2023331885402009779
The new chonky Qwen 3.5 looks pretty solid, beating their own Qwen3-Max model everywhere and is much better at vision benchmarks than Qwen3-235B-A22B-VL Now what I sadly haven’t seen is anything on reasoning efficiency.”” https://x.com/scaling01/status/2023343368399704506
Kimi K2‑0905 and Qwen3‑Max preview: two 1T open weights models launched | AINews https://news.smol.ai/issues/25-09-05-1t-models
Multilingual data finally moves from “data collection” to “data curation”. UberWeb breaches the compute-performance frontier for multilingual data. We spell out all our learnings from this year long effort at training multilingual models at the 20 trillion token data scale.”” https://x.com/pratyushmaini/status/2024157352862376280
Leave a Reply