{"id":"2028500277796720739","url":"https://x.com/cgtwts/status/2028500277796720739","text":"Qwen 3.5 Small models \n\n- fully open source\n- beats models 4x it’s size\n- 9B model performs on par with GPT OSS 120B while being 13x smaller\n- outperforms Gemini 3 flash and Claude sonnet 4.5 on select benchmarks \n- runs on any laptop\n- even works on a phone \n- completely free.","author":{"name":"CG","username":"cgtwts","avatarUrl":"https://pbs.twimg.com/profile_images/1955738644528160769/cgVyKTcs_200x200.jpg"},"createdAt":"Mon Mar 02 15:58:48 +0000 2026","engagement":{"replies":124,"retweets":447,"likes":6847,"views":656487},"media":{"photos":[],"videos":[{"url":"https://video.twimg.com/amplify_video/2028500216979279873/vid/avc1/1280x720/0h3eGZfTnWem9pTD.mp4?tag=21","thumbnailUrl":"https://pbs.twimg.com/amplify_video_thumb/2028500216979279873/img/PMnm_8ZLiK6z1sou.jpg","width":1280,"height":720,"duration":17.297}]},"quoteTweet":{"id":"2028460046510965160","url":"https://x.com/Alibaba_Qwen/status/2028460046510965160","text":"🚀 Introducing the Qwen 3.5 Small Model Series\nQwen3.5-0.8B · Qwen3.5-2B · Qwen3.5-4B · Qwen3.5-9B\n\n✨ More intelligence, less compute.\nThese small models are built on the same Qwen3.5 foundation — native multimodal, improved architecture, scaled RL:\n• 0.8B / 2B → tiny, fast, great for edge device\n• 4B → a surprisingly strong multimodal base for lightweight agents\n• 9B → compact, but already closing the gap with much larger models\nAnd yes — we’re also releasing the Base models as well.\nWe hope this better supports research, experimentation, and real-world industrial innovation.\nHugging Face: https://huggingface.co/collections/Qwen/qwen35\nModelScope: https://modelscope.cn/collections/Qwen/Qwen35","author":{"name":"Qwen","username":"Alibaba_Qwen","avatarUrl":"https://pbs.twimg.com/profile_images/2064231947149377536/Ab70PxT5_200x200.jpg"},"createdAt":"Mon Mar 02 13:18:56 +0000 2026"}}