Nemotron Cascade 2 30B A3B
nvidia/Nemotron-Cascade-2-30B-A3B
Released2026-03-19Size30B · 3B activeArchitecturehybrid linear attnLicensenvidia-open-model-licenseAA rank#53 of 75
Files and memory at 8K context
| Quant | File | Files | + KV + runtime | Smallest fit | Source |
|---|---|---|---|---|---|
| Q4_K_M mradermacher/Nemotron-Cascade-2-30B-A3B-GGUF | 22.83 GiB | 1 | 23.9 GiB+ | 32 GB GPU | HF files |
| Q5_K_M mradermacher/Nemotron-Cascade-2-30B-A3B-GGUF | 24.25 GiB | 1 | 25.3 GiB+ | 32 GB GPU | HF files |
| Q8_0 mradermacher/Nemotron-Cascade-2-30B-A3B-GGUF | 31.28 GiB | 1 | 32.3 GiB+ | 48 GB GPU | HF files |
| FP16 freddm/Nemotron-Cascade-2-30B-A3B-GGUF | 58.84 GiB | 1 | 59.9 GiB+ | 64 GB GPU | HF files |
| MLX 4-bit mlx-community/Nemotron-Cascade-2-30B-A3B-4bit | 16.55 GiB | 4 | 17.6 GiB+ | 32 GB Mac | HF files |
| MLX 8-bit mlx-community/Nemotron-Cascade-2-30B-A3B-8bit | 31.26 GiB | 7 | 32.3 GiB+ | 48 GB Mac | HF files |
Sizes are exact file bytes (text weights; projectors listed separately). “+” = recurrent state or other terms unknown. Smallest fit adds 0.5 GiB spare; Macs use 75% of memory.
Quality
Artificial Analysis Intelligence Index
11.7 AA estimate
Variant: Nemotron Cascade 2 30B A3B · observed 2026-10-17 · source
Rank #53 of 75 researched 2026 models.
Arena text rating
Unknown
Scores describe the evaluated hosted variant, not each quantized file. AA and Arena are never averaged.
Reported speeds 0
No reported measurement yet. Submit one.
Community threads
- Nemotron Cascade mutates nested path literals in tool arguments even with explicit literal-preservation instructions and temperature=0. Hugging Face · 2026-03-31
- modeling_nemotron_h.py: Multiple bugs in HybridMambaAttentionDynamicCache break generation with CUDA fast path Hugging Face · 2026-03-25
Qualitative signals only; titles are not verified claims.