What GPU do I need to run dphn/Dolphin3.0-R1-Mistral-24B?
A 23.6B model tuned to reason step by step before answering. 23.6B parameters, published in BF16. View on Hugging Face
Dolphin3.0-R1-Mistral-24B is published by dphn on Hugging Face, with 1,098 downloads and 228 likes to date. It's a MistralForCausalLM model built for text-generation, published natively in BF16.
What Dolphin3.0-R1-Mistral-24B is
Dolphin3.0-R1-Mistral-24B is a 23.6B-parameter language model published by Cognitive Computations (Dolphin) on Hugging Face. It works through a problem step by step before answering, rather than responding directly. It is released under Not stated.
License note: no license tag published on the model's Hugging Face card; check the repo directly before any commercial use. Facts in this section are sourced from Dolphin3.0-R1-Mistral-24B's Hugging Face model card, not benchmarked by Aquanode.
What it's used for
- Multi-step reasoning and math
- Code generation
- Agentic tool use
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 | 43.9 GB | 52.7 GB | A100 | 1 | $1.16/hr |
| cheaper alt. | RTX A5000 | 3 | $0.528/hr | ||
| FP8 (quantized) | 22.0 GB | 26.3 GB | RTX 5000 Ada | 1 | $0.539/hr |
| cheaper alt. | RTX 4070 Super | 3 | $0.363/hr | ||
| INT4 (quantized) | 11.0 GB | 13.2 GB | RTX A4000 | 1 | $0.167/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 Dolphin3.0-R1-Mistral-24B at its published (BF16) precision: 1× A100, at $1.16/hr per GPU ($1.16/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Dolphin3.0-R1-Mistral-24B: 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: 163,840 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.63 GB |
| 32K | 5.0 GB |
Computed from the layer, head and window counts in the model's own configuration (dphn/Dolphin3.0-R1-Mistral-24B), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
Dolphin3.0-R1-Mistral-24B: common questions
Can Dolphin3.0-R1-Mistral-24B run on a single GPU?
Yes, but not on a desktop card. At BF16 it needs 52.7 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest live single-card fit is a 80.0 GB A100 at $1.16/hr.
What is the least VRAM Dolphin3.0-R1-Mistral-24B can run in?
13.2 GB, at INT4 (quantized), which fits a 16 GB card, against 52.7 GB at BF16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Does quantizing Dolphin3.0-R1-Mistral-24B lower the GPU bill?
Yes. At BF16 the cheapest live fit is one A100 at $1.16/hr. At INT4 (quantized) it drops to one RTX A4000 at $0.167/hr, provided a quantized checkpoint exists for it.
How to run Dolphin3.0-R1-Mistral-24B
Run Dolphin3.0-R1-Mistral-24B with vLLM
Generic example, not from the model's own docs: adjust flags (quantization, context length, parallelism) for your setup.
vllm serve dphn/Dolphin3.0-R1-Mistral-24B --tensor-parallel-size 1Run Dolphin3.0-R1-Mistral-24B with GGUF quantizations
Prebuilt GGUF weights published at mradermacher/Dolphin3.0-R1-Mistral-24B-GGUF. Run with llama.cpp's llama-server or load the repo directly in LM Studio.
llama-server -hf mradermacher/Dolphin3.0-R1-Mistral-24B-GGUFSource: https://huggingface.co/mradermacher/Dolphin3.0-R1-Mistral-24B-GGUF
Deploy Dolphin3.0-R1-Mistral-24B on Aquanode
Aquanode has no one-click deploy template for Dolphin3.0-R1-Mistral-24B; you install the inference engine yourself with the commands below. Aquanode sells GPU pods billed per second, not a hosted inference API.
- Launch a bare GPU pod sized to the requirement above (1× A100 or larger).
- Open a terminal on the pod, or save one of the commands above as a startup script so it runs automatically the first time the pod boots.
- Run the command and connect to the resulting endpoint.
Launch a GPU podWeight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More Mistral Small models
- Mistral-Small-24B-Instruct-2501 (23.6B, BF16)
- Dolphin-Mistral-24B-Venice-Edition (24.0B, BF16)
Alternatives at this size
Other models for text-generation within about a third of Dolphin3.0-R1-Mistral-24B's 23.6B parameters, from other model lines.
- Qwen3-32B (32.8B, BF16)
- GLM-4.7-Flash (31.2B, BF16)
- NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (31.6B, BF16)
- gpt-neox-20b (20.7B, F16)
- granite-4.1-30b (28.9B, BF16)
More on Dolphin3.0-R1-Mistral-24B
Fits on an 80 GB GPU at BF16: every model that fits in 80 GB.
Fits on a 32 GB GPU at FP8: every model that fits in 32 GB.
Fits on a 16 GB GPU at INT4: every model that fits in 16 GB.
How to run Mistral Small and how to fine-tune Mistral Small.
Related reading: A100 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, Best GPU for LLM inference, and What AI model fine-tuning is.
