LLM EXPLORER 67,952 MODELS INDEXED

Qwen3.8 Flash Next OQ3.5e Fp16 Mtp by Robot-Haus

By Robot-Haus · 766 downloads

Qwen3.8 Flash Next OQ3.5e Fp16 Mtp is an open-source language model by Robot-Haus. Features: 180b LLM, VRAM: 90.7GB, License: other, Quantized, LLM Explorer Score: 0.28.

  Arxiv:2502.02789   3-bit   3.5-bit   Apple-silicon Base model:quantized:qwen/qwen... Base model:qwen/qwen3.8-flash-...   Conversational Dataset:ai-safety-promise/safe...   Dataset:cais/mmlu Dataset:google-research-datase... Dataset:huggingfaceh4/hellaswa... Dataset:huggingfaceh4/truthful... Dataset:livecodebench/livecode... Dataset:openai/openai humaneva...   En   Fp16   Image-text-to-text   Mixture-of-experts   Mlx   Mlx-vlm   Model-index   Mtp   Omlx   Oq   Oqe   Quantized   Qwen   Qwen3.8   Qwen4-exp   Qwen4 exp   Region:us   Safetensors   Sharded   Speculative-decoding   Tensorflow   Vision   Vision-language

Qwen3.8 Flash Next OQ3.5e Fp16 Mtp Parameters and Internals

LLM NameQwen3.8 Flash Next OQ3.5e Fp16 Mtp
Repository πŸ€—https://huggingface.co/Robot-Haus/Qwen3.8-Flash-Next-oQ3.5e-fp16-mtp 
Base Model(s)  Qwen/Qwen3.8-Flash-Next   Qwen/Qwen3.8-Flash-Next
Model Size180b
Required VRAM90.7 GB
Updated2026-10-04
MaintainerRobot-Haus
Model Typeqwen4_exp
Model Files  5.1 GB: 1-of-18   5.0 GB: 2-of-18   5.0 GB: 3-of-18   5.0 GB: 4-of-18   5.0 GB: 5-of-18   5.6 GB: 6-of-18   5.0 GB: 7-of-18   5.3 GB: 8-of-18   5.2 GB: 9-of-18   5.2 GB: 10-of-18   5.3 GB: 11-of-18   5.2 GB: 12-of-18   5.2 GB: 13-of-18   5.3 GB: 14-of-18   5.2 GB: 15-of-18   5.2 GB: 16-of-18   5.3 GB: 17-of-18   2.6 GB: 18-of-18
Supported Languagesen
Quantization Typefp16
Model ArchitectureQwen4ExpForConditionalGeneration
Licenseother
Model Max Length262144
Transformers Version5.8.0.dev0
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Errorsreplace

Best Alternatives to Qwen3.8 Flash Next OQ3.5e Fp16 Mtp

Best Alternatives
Context / RAM
Downloads
Likes
...wen3.8 Flash Next MLX 4bit MTP0K / 112.8 GB446012
...wen3.8 Flash Next MLX 4bit MTP0K / 112.8 GB16435
...en3.8 Flash Next OQ4e Fp16 Mtp0K / 107.6 GB3872
...en3.8 Flash Next OQ4e Fp16 Mtp0K / 107.3 GB7462
...n3.8 Flash Next GSQ RCO Q2 Mlx0K / 4.8 GB4452
Qwen3.8 Flash Next 4bit Mtp0K / 112.9 GB5403
...3.8 Flash Next OQ2.5e Fp16 Mtp0K / 83.6 GB2133
...3.8 Flash Next OQ3.5e Fp16 Mtp0K / 97.2 GB1451
...en3.8 Flash Next OQ3e Fp16 Mtp0K / 91.8 GB1321
...MLX Activation 3bit Native PLE0K / 34.7 GB2841
Note: green Score (e.g. "73.2") means that the model is better than Robot-Haus/Qwen3.8-Flash-Next-oQ3.5e-fp16-mtp.