Models / Qwen3-Nextgenerated from capability-matrix.json
Qwen3-Next
Qwen's 80B mixture-of-experts with about 3B active per token, on hybrid attention.
chatexperts: sparse +sharedsafetensors
Get it
goinfer can't fetch this one yet. Download the safetensors folder with your usual tool, then point goinfer at it.
goinfer-serve -model ./path/to/checkpoint-folderWhat hasn't been shown
- goinfer can't download it for you: this family only loads from safetensors, which come as several files. That's planned (checkpoint fetch, P1–P9).
- No vetted checkpoint, so there's no fit verdict, no speed and no tool-calling result for this family.
How sure we are
Against the released model100.0%
picks the same next token as HuggingFace
0.98931
closest the raw scores get at their worst position (cosine, 1.00000 is identical; bar starts at 0.95)
real-oracle 100.0%/0.98931 · what parity-gated means
Measured speed · decode, tokens per second
not measured — no vetted checkpoint to measure
Architecture, for the curious
- model_type
- qwen3_next
- design
- gated-linear hybrid (Gated DeltaNet)
- experts
- sparse +shared
- attention window
- none
- QK-norm
- yes
- RoPE
- partial
- norm
- RMSNorm, pre-norm
- activation
- SwiGLU
- tied head
- no
- modality
- text
- GPU-resident
- eligible
registry description: Qwen3-Next 80B-A3B: same DeltaNet/softmax/MoE hybrid shape as Qwen3.5, computed (not stated) layer pattern