GLM-4.5-Air vs NVIDIA-Nemotron-3-Super-120B-A12B-BF16
GLM-4.5-Air (110.5B parameters) and NVIDIA-Nemotron-3-Super-120B-A12B-BF16 (123.6B parameters) side by side: the memory each needs at every precision, what it costs to run on a live GPU, and the context window, KV cache and license where they are published. Numbers are computed from the models' published specs; this page does not rank quality.
Side by side
| Fact | GLM-4.5-Air | NVIDIA-Nemotron-3-Super-120B-A12B-BF16 |
|---|---|---|
| Parameters | 110.5B (Mixture-of-experts: 8 of 128 experts active per token (exact active-parameter count not stated on the model card)) | 123.6B |
| Architecture | Grouped-query attention; mixture of 128 experts, 8 active per token | Hybrid (some layers use full attention); mixture of 512 experts, 22 active per token |
| Context length | 128K tokens (131,072) | 262,144 tokens |
| License | MIT | – |
| Published precision | BF16 | BF16 |
| VRAM needed, As published | 247 GB | 276 GB |
| VRAM needed, FP8 | 123 GB | 138 GB |
| VRAM needed, INT4 | 61.7 GB | 69.1 GB |
| Cheapest live fit, As published | RTX A6000 × 6 · $2.18/hr | RTX A6000 × 6 · $2.18/hr |
| Cheapest live fit, FP8 | RTX 4080 Super × 4 · $1.35/hr | RTX 4000 SFF Ada × 7 · $1.39/hr |
| Cheapest live fit, INT4 | A100 · $1.21/hr | A100 · $1.21/hr |
| KV cache per token (16-bit) | 184 KB | 8 KB |
| KV cache at 32k tokens | 5.75 GB | 0.25 GB |
| KV cache at 128k tokens | 23.0 GB | 1.00 GB |
VRAM is the weight size at each precision times a flat 1.2 overhead; see the methodology. The FP8 and INT4 rows need a quantized checkpoint or an engine that quantizes on load. The fit is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does. KV cache is for one sequence at 16-bit, computed from each model's config where the attention layout is known.
Which to pick
- GLM-4.5-Air needs less VRAM at its published precision (247 GB against 276 GB), so it fits on a smaller GPU.
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 lists the longer context window (262,144 tokens against 131,072).
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 caches less per sequence at 32k tokens (0.3 GB against 5.8 GB), leaving more memory for batching.
These follow only from the facts in the table above. Whether either model does your task well is a separate question this page does not answer.
Keep reading
- GLM-4.5-Air: full VRAM table and live GPU fit
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16: full VRAM table and live GPU fit
- The GLM model series
- The Nemotron model series
- All models that fit in 288 GB
Other comparisons
- GLM-4.5-Air vs Qwen3-Next-80B-A3B-Instruct
- GLM-4.5-Air vs Hunyuan-A13B-Instruct
- GLM-4.5-Air vs Qwen3-Next-80B-A3B-Thinking
- GLM-4.5-Air vs NVIDIA-Nemotron-3-Super-120B-A12B-Base-BF16
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 vs Qwen3-Next-80B-A3B-Instruct
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 vs Hunyuan-A13B-Instruct
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 vs Qwen3-Next-80B-A3B-Thinking
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 vs GLM-4.5-Air-Base
