Qwen3.6 VL REAP 26B A3B is an open-source language model by atbender. Features: 26.6b LLM, VRAM: 53.2GB, Context: 256K, License: apache-2.0, MoE.
Arxiv:2503.02138 Agentic-coding Base model:finetune:qwen/qwen3... Base model:qwen/qwen3.6-35b-a3... Conversational Endpoints compatible Mixture-of-experts Moe Multimodal Pruned Qwen3.6 Qwen3 5 moe text Reap Region:us Safetensors Sharded Tensorflow Vision Vision-language
Model Card on HF 🤗: https://huggingface.co/atbender/Qwen3.6-VL-REAP-26B-A3B
Qwen3.6 VL REAP 26B A3B Parameters and Internals
| LLM Name | Qwen3.6 VL REAP 26B A3B |
| Repository 🤗 | https://huggingface.co/atbender/Qwen3.6-VL-REAP-26B-A3B |
| Base Model(s) | |
| Model Size | 26.6b |
| Required VRAM | 53.2 GB |
| Updated | 2026-04-20 |
| Maintainer | atbender |
| Model Type | qwen3_5_moe_text |
| Model Files | |
| Model Architecture | Qwen3_5MoeForCausalLM |
| License | apache-2.0 |
| Context Length | 262144 |
| Model Max Length | 262144 |
| Transformers Version | 5.6.0.dev0 |
| Tokenizer Class | TokenizersBackend |
| Padding Token | <|endoftext|> |
| Vocabulary Size | 248320 |
| Errors | replace |
Rank the Qwen3.6 VL REAP 26B A3B Capabilities
🆘 Have you tried this model? Rate its performance. This feedback would greatly assist ML community in identifying the most suitable model for their needs. Your contribution really does make a difference! 🌟
What open-source LLMs or SLMs are you in search of? 53026 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Check out Ag3ntum — our secure, self-hosted AI agent for server management.
Release v20260328a