What GPU do I need to run inclusionAI/Ling-3.0-flash?
127.5B parameters, published in BF16. View on Hugging Face
Ling-3.0-flash is published by inclusionAI on Hugging Face, with 20,206 downloads and 385 likes to date. It's a BailingMoeV3ForCausalLM model built for text-generation, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 237.5 GB | 285.0 GB | RTX A6000 | 6 | $3.63/hr |
| FP8 (quantized) | 118.7 GB | 142.5 GB | RTX 4000 Ada | 8 | $2.12/hr |
| INT4 (quantized) | 59.4 GB | 71.2 GB | A100 | 1 | $1.37/hr |
| cheaper alt. | RTX 3090 | 3 | $0.660/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 Ling-3.0-flash at its published (BF16) precision: 6× RTX A6000, at $0.605/hr per GPU ($3.63/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Ling-3.0-flash: 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.
Multi-head latent attention: every layer caches one compressed latent per token. Cached per token: 48,384 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.18 GB |
| 32K | 1.5 GB |
| 128K | 5.9 GB |
| 256K (model maximum) | 11.8 GB |
Computed from the layer, head and window counts in the model's own configuration (inclusionAI/Ling-3.0-flash), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
Ling-3.0-flash: common questions
Can Ling-3.0-flash run on a single GPU?
No. At BF16 it needs 285.0 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest capable card in the live feed is a 48.0 GB RTX A6000, and it takes 6 of them.
How many GPUs do I need to run Ling-3.0-flash?
6 at BF16. It needs 285.0 GB of VRAM and the cheapest capable live offer is a 48.0 GB RTX A6000, so 6 of them come to $3.63/hr in total.
Does quantizing Ling-3.0-flash lower the GPU bill?
Yes. At BF16 the cheapest live fit is 6 RTX A6000 cards at $3.63/hr. At INT4 (quantized) it drops to one A100 at $1.37/hr, provided a quantized checkpoint exists for it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More Ling models
- Ling-3.0-tiny (7.9B, BF16)
- Ling-mini-2.0 (16.3B, BF16)
Alternatives at this size
Other models for text-generation within about a third of Ling-3.0-flash's 127.5B parameters, from other model lines.
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16 (123.6B, BF16)
- GLM-4.5-Air (110.5B, BF16)
- Laguna-S-2.1 (117.6B, BF16)
- xLAM-8x22b-r (140.6B, BF16)
- bloom (176.2B, BF16)
More on Ling-3.0-flash
Fits on a 288 GB GPU at BF16: every model that fits in 288 GB.
Fits on a 192 GB GPU at FP8: every model that fits in 192 GB.
Fits on an 80 GB GPU at INT4: every model that fits in 80 GB.
Best chat and assistants models: how Ling-3.0-flash ranks against the rest.
Related reading: RTX A6000 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.
Your own GPU cloud accounts
One console for the GPU clouds you already rent from. Your keys stay on your machine.
$20 a month, flat