What GPU do I need to run zai-org/GLM-ASR-Nano-2512?
2.3B parameters, published in BF16. View on Hugging Face
Set up GLM-ASR-Nano-2512GLM-ASR-Nano-2512 is published by zai-org on Hugging Face, with 79,017 downloads and 389 likes to date. It's a GlmAsrForConditionalGeneration model built for automatic-speech-recognition, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for activations and allocator fragmentation. Speech models don't build the same growing KV-cache a text model does. Memory scales primarily with input audio length. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 4.2 GB | 5.0 GB | RTX 3060 | 1 | $0.110/hr |
| FP8 (quantized) | 2.1 GB | 2.5 GB | RTX 5060 Ti | 1 | $0.188/hr |
| INT4 (quantized) | 1.1 GB | 1.3 GB | RTX 3060 | 1 | $0.110/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run GLM-ASR-Nano-2512 at its published (BF16) precision: 1× RTX 3060, at $0.110/hr per GPU ($0.110/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
GLM-ASR-Nano-2512: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
Grouped-query attention: every layer caches keys and values for a few shared KV heads. Cached per token: 57,344 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.22 GB |
| 8K (model maximum) | 0.44 GB |
Computed from the layer, head and window counts in the model's own configuration (zai-org/GLM-ASR-Nano-2512), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
GLM-ASR-Nano-2512: common questions
How much VRAM does GLM-ASR-Nano-2512 need?
5.0 GB at BF16, 2.5 GB at FP8 (quantized), 1.3 GB at INT4 (quantized). All of those sit under 6 GB, the smallest capacity this page reasons about, so VRAM is not what limits where this model runs. The 4.2 GB of weights plus inference overhead is the whole requirement.
How many copies of GLM-ASR-Nano-2512 fit on one RTX 3060?
2, by VRAM alone. That card carries 12.0 GB and one copy needs 5.0 GB at BF16, on a live rate of $0.110/hr for the whole card. Throughput is not modelled here, so 2 copies is not 2 times the requests served.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More zai-org models
- GLM-OCR (1.3B, BF16)
Alternatives at this size
Other models for automatic-speech-recognition within about a third of GLM-ASR-Nano-2512's 2.3B parameters, from other model lines.
- whisper-large-v3 (1.5B, F16)
- Qwen3-ASR-1.7B (2.3B, BF16)
- cohere-transcribe-03-2026 (2.1B, BF16)
- seamless-m4t-v2-large (2.3B, F32)
- granite-speech-4.1-2b (2.3B, BF16)
More on GLM-ASR-Nano-2512
Fits on an 8 GB GPU at BF16: every model that fits in 8 GB.
Best speech-to-text models: how GLM-ASR-Nano-2512 ranks against the rest.
Related reading: H100 pricing and specs, Whisper variants compared, and What AI inference is.
