minjaechoi/qwen3p6-35b-a3b-2p04bit-r73
TEXT GENERATIONPricing:Input $0.4 / Cached $0.07 / Output $4Concurrent Unit Cost:2Model Size:35.1BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Oct 7, 2026Architecture:Transformer Featherless Exclusive Cold
The minjaechoi/qwen3p6-35b-a3b-2p04bit-r73 is a 35.1 billion parameter language model based on the Qwen3.6-35B-A3B architecture. This model features routed experts with an average quantization of 2.0350 bits, while other weights remain in BF16 format. It is designed for efficient loading with standard `transformers` and vLLM libraries, making it suitable for applications requiring reduced memory footprint and potentially faster inference due to quantization.
Loading preview...
Overview
This model, minjaechoi/qwen3p6-35b-a3b-2p04bit-r73, is an internal research checkpoint derived from the Qwen/Qwen3.6-35B-A3B base model. It incorporates a unique quantization strategy to optimize its footprint and performance.
Key Characteristics
- Base Model: Built upon the Qwen3.6-35B-A3B architecture.
- Quantization: Utilizes "routed experts" where the average weight precision is 2.0350 bits. All other weights within the model are maintained in BF16 (bfloat16) format.
- Loading Compatibility: Designed to be loaded seamlessly using standard
transformersand vLLM libraries, ensuring ease of integration into existing workflows. - Internal ID: Identified internally as
r73.
Good for
- Experimentation with highly quantized models for efficiency.
- Scenarios where memory footprint reduction is critical.
- Research into mixed-precision model deployment and performance.