myvio/qwen3-next-80b • LM Studio Hubqwen3-next-80b
qwen3-next-80b
Description
Hybrid attention architecture, high-sparsity Mixture-of-Experts 80B model (active 3B). Currently supported for Mac only with MLX.
Minimum system memory
42GB
Description
Hybrid attention architecture, high-sparsity Mixture-of-Experts 80B model (active 3B). Currently supported for Mac only with MLX.
Minimum system memory
42GB
Description
Hybrid attention architecture, high-sparsity Mixture-of-Experts 80B model (active 3B). Currently supported for Mac only with MLX.
Minimum system memory
42GB
Description
Hybrid attention architecture, high-sparsity Mixture-of-Experts 80B model (active 3B). Currently supported for Mac only with MLX.
Minimum system memory
42GB
Qwen3 Next 80B
The first model in the Qwen3-Next series featuring innovative hybrid attention architecture and high-efficiency Mixture-of-Experts design.
Key Features
Parameters
Custom configuration options included with this model
Sources
The underlying model files this model uses
Qwen3 Next 80B
The first model in the Qwen3-Next series featuring innovative hybrid attention architecture and high-efficiency Mixture-of-Experts design.
Key Features
Parameters
Custom configuration options included with this model
Sources
The underlying model files this model uses
- 80B total parameters, 3B activated (A3B)
- 48 layers with hybrid layout
- 512 experts with only 10 activated per token
- Context length: 262,144 tokens
- No thinking mode support (instruct-only)
Delivers performance comparable to much larger models while maintaining exceptional efficiency:
- Outperforms Qwen3-32B with 10x inference throughput for long contexts
- Matches Qwen3-235B-A22B on many benchmarks with significantly lower computational requirements
- Superior ultra-long-context handling up to 256K+ tokens
- 80B total parameters, 3B activated (A3B)
- 48 layers with hybrid layout
- 512 experts with only 10 activated per token
- Context length: 262,144 tokens
- No thinking mode support (instruct-only)
Delivers performance comparable to much larger models while maintaining exceptional efficiency:
- Outperforms Qwen3-32B with 10x inference throughput for long contexts
- Matches Qwen3-235B-A22B on many benchmarks with significantly lower computational requirements
- Superior ultra-long-context handling up to 256K+ tokens
