Arxiv:1910.09700 Autotrain compatible Conversational Endpoints compatible Moe Qwen3 moe Region:us Safetensors
Model Card on HF 🤗: https://huggingface.co/leeminwaan/Qwen3-MOE-4x0.6B-2.4B-reasoning-v1-full
Qwen3 MoE 4x0.6B 2.4B Reasoning V1 Full Benchmarks
nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").
What is the LLM Explorer Rank (Score)
Qwen3 MoE 4x0.6B 2.4B Reasoning V1 Full Parameters and Internals
| LLM Name | Qwen3 MoE 4x0.6B 2.4B Reasoning V1 Full |
| Repository 🤗 | https://huggingface.co/leeminwaan/Qwen3-MOE-4x0.6B-2.4B-reasoning-v1-full |
| Model Size | 1.4b |
| Required VRAM | 2.8 GB |
| Updated | 2025-09-21 |
| Maintainer | leeminwaan |
| Model Type | qwen3_moe |
| Model Files | |
| Model Architecture | Qwen3MoeForCausalLM |
| Context Length | 40960 |
| Model Max Length | 40960 |
| Transformers Version | 4.52.4 |
| Tokenizer Class | Qwen2Tokenizer |
| Padding Token | <|im_end|> |
| Vocabulary Size | 151936 |
| Torch Data Type | float16 |
| Errors | replace |
Rank the Qwen3 MoE 4x0.6B 2.4B Reasoning V1 Full 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? 51507 in total.
Original data from HuggingFace, OpenCompass and various public git repos.
Release v20241124