Qwen3.5 4B
alibaba/qwen3.5-4b- Alibaba
- Family · Qwen3
- Released 2 Mar 2026
Not on OSMEXO yet
Scores · standing
Intelligence13.1#194/489
Coding22.6#134/196
News & trends
23 mentions this week ▲ 35%this model · videos, posts, articles · 30 days
Trending · 14 days
- Every Neuron in an LLM (Qwen3.5 4B)1.6K views
- Interactive neuron steering playground for Qwen3.5-4B that runs locally in the browser on WebGPU216 upvotes
- Neuron Explorer for Qwen3.5-4B: See every neuron at work, make targeted edits26 points
- Spark X2.6 4B Local Test | Beats Qwen3.5 4B? | Agentic Work, Coding & RAG Test | 16GB Local AI Setup3.1K views
- Top 10 AI agents for small local LLMs in 2026, tested on Qwen3.5-4B146 upvotes
News · latest from Qwen3
- Memdoor: A team of 5 Qwen3.8-27B models beats Fable 5 on LiveCodeBench Hardgithub.com · 31 Oct
- Neuron Explorer for Qwen3.5-4B: See every neuron at work, make targeted editsgithub.com · 26 Oct
- How to train your own Jev for $17Together AI · 23 Sep
Benchmarks
bar = share of the top score · Benchmark data: Artificial AnalysisSpeed · median across providers, latest day
Output speed#100/10120 tok/s5.9× slower than median
First token#240.53 s1.8× quicker than median
Family progress
Intelligence Index
- Family releases33.7
Research papers
- Low-Bit GDN Recurrent-State Quantization: Global Gradient-aware Allocation and Pure-Write Sliding Windows29 Oct
- Look at Where You Think: Localization-Head Alignment for Fine-Grained Perception in Thinking VLMs29 Oct
- Weaken to Strengthen: Weakening-Induced On-Policy Self-Distillation for Large Reasoning Models28 Oct
- MLE-World: Predicting Machine Learning Experiment Outcomes with a Learned World Model27 Oct
- ThinkJev: One Thought, Every Budget for Calibrated Typed Decisions27 Oct
- Back to Basics: A Skill-Centric Approach to Generalist Web Agents26 Oct
- MiRL: Error-Targeted Reinforcement Learning for Correcting Terminal Agent Misbehaviors26 Oct
- RIFT: KV Cache Compression for Draft Models with Complementary Compressibility Training22 Oct
Social
Most watched and discussed · 90 days
- Spark X2.6 4B Local Test | Beats Qwen3.5 4B? | Agentic Work, Coding & RAG Test | 16GB Local AI SetupVenelin Valkov · 3.1K views · 21 Oct
- Every Neuron in an LLM (Qwen3.5 4B)plusgiant5 · 1.6K views · 30 Oct
- 小米开源 Xiaomi-Lens-0！基于 Qwen3.5-4B，支持目标定位、实例分割与细粒度视觉理解鲲鹏Talk · 338 views · 22 Oct
- Qwen3.5 4B at 8-Bit: Half the Speed on a 16 GB Mac?Inside The Machine · 332 views · 28 Oct
- 【Qwenputer】Qwen3.5 4B の重みに RISC-V の CPU を埋めたモデルを手元で動かしたら、大きな掛け算 40 問中 30 問を正解 (素の Qwen は 1 問)さつきのOSS研究室 · 127 views · 29 Oct
- Can One Neuron Make an AI Hostile? We Tested It on Qwen3.5-4BShashank · 124 views · 26 Oct
- 【neuron-explorer】Qwen3.5 4Bの1個のニューロンを押して試したら自己紹介がエイリアンに変わった #ShortsさつきのOSS研究室 · 85 views · 30 Oct
- Qwen3.5 4B 3D Coder: 4.5B, Apache 2.0, Three.jsAIQuirk · 55 views · 31 Oct
- 【neuron-explorer】Qwen3.5 4B の 1 個のニューロンを押したら自己紹介がエイリアンに変わった｜手元で実測し作者の記録と 1 層 9,216 個全部で一致さつきのOSS研究室 · 54 views · 30 Oct
- Qwen3.5 4B 3D Coder + 3 more models · hotin.ai daily Oct 28, 2026 #ShortsHotInAI · 39 views · 28 Oct
- Qwen3.5 4B at 4-Bit vs 8-Bit on a 16 GB Mac: What It Really CostsInside The Machine · 27 views · 28 Oct
- Qwenputer: A RISC-V CPU Compiled Inside Qwen3.5-4B's WeightsSadra Majidi · 4 views · 31 Oct
Same family
Qwen3- 6.0
- 7.2
- 6.9
- 6.1
- 5.2
- 9.5
Alternatives
other vendors · similar capability- 12.9
- 8.5
- 11.6
- 13.1
Want Qwen3.5 4B on OSMEXO?
Sign in to request it. Requests are reviewed; a request does not guarantee availability.