PicoMistral 23M — Hardware Requirements & GPU Compatibility
Chat
PicoMistral 23M is a 24M-parameter open language model from PicoKittens in the Mistral family. It supports a context window of up to 512 tokens. At BF16 it needs about 0.36 GB of VRAM — see which GPUs and Macs can run it below.
1.1K downloads
10 likesSpecifications
- Publisher
- PicoKittens
- Family
- Mistral
- Parameters
- 24M
- Architecture
- MistralForCausalLM
- Context Length
- 512 tokens
- Vocabulary Size
- 16,384
- Release Date
- 2026-03-03
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does PicoMistral 23M Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| BF16 | 16.00 | 0.4 GB | — | 0.05 GB | Brain floating point 16 — preferred for training |
Which GPUs Can Run PicoMistral 23M?
BF16 · 0.4 GBPicoMistral 23M (BF16) requires 0.4 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 1+ GB is recommended. 35 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti.
Runs great
— Plenty of headroomWhich Devices Can Run PicoMistral 23M?
BF16 · 0.4 GB33 devices with unified memory can run PicoMistral 23M, including NVIDIA DGX H100, NVIDIA DGX A100 640GB.
Runs great
— Plenty of headroomRelated Models
Frequently Asked Questions
- How much VRAM does PicoMistral 23M need?
PicoMistral 23M requires 0.4 GB of VRAM at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 24M × 16 bits ÷ 8 = 0 GB
KV Cache + Overhead ≈ 0.4 GB (at 2K context + ~0.3 GB framework)
VRAM usage by quantization
BF160.4 GB- Can I run PicoMistral 23M on a Mac?
PicoMistral 23M requires at least 0.4 GB at BF16, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run PicoMistral 23M locally?
Yes — PicoMistral 23M can run locally on consumer hardware. At BF16 quantization it needs 0.4 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is PicoMistral 23M?
At BF16, PicoMistral 23M can reach ~8097 tok/s on AMD Instinct MI300X. On NVIDIA GeForce RTX 4090: ~1820 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 0.4 × 0.55 = ~8097 tok/s
Estimated speed at BF16 (0.4 GB)
~8097 tok/s~1820 tok/s~6052 tok/s~5006 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of PicoMistral 23M?
At BF16, the download is about 0.05 GB.
- Which GPUs can run PicoMistral 23M?
35 consumer GPUs can run PicoMistral 23M at BF16 (0.4 GB). Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 35 GPUs have plenty of headroom for comfortable inference.
- Which devices can run PicoMistral 23M?
33 devices with unified memory can run PicoMistral 23M at BF16 (0.4 GB), including Mac Mini M4 (16 GB), Mac Mini M4 (32 GB), Mac Mini M4 Pro (24 GB), Mac Mini M4 Pro (48 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.