LLMLocal inference
Codestral 22B
VRAM requirements to run Codestral 22B locally at each quantization level. Find the cheapest GPU that fits below.
Q4 VRAM
13 GB
Q8 VRAM
24 GB
FP16 VRAM
44 GB
Context window
32 k tokens
01 // GPUs that can run Codestral 22B
Cheapest compatible hardware by quantization
Sorted cheapest first. All prices are approximate street prices.
Q4Q4_K_M (4-bit)
needs ≥13 GB VRAMGPUVRAMPriceTier
Q8Q8_0 (8-bit)
needs ≥24 GB VRAMGPUVRAMPriceTier
FP16FP16 (full precision)
needs ≥44 GB VRAMGPUVRAMPriceTier
02 // Frequently asked
Codestral 22B GPU questions
How much VRAM does Codestral 22B need?
Codestral 22B requires approximately 13GB VRAM at Q4 quantization, 24GB at Q8, or 44GB at full FP16 precision. Q4 is the most practical choice for consumer hardware.
What is the cheapest GPU to run Codestral 22B?
The cheapest single GPU that fits Codestral 22B at Q4 is the Radeon RX 9070 XT (16GB VRAM, ~$549). At Q4 you need at least 13GB.
Can I run Codestral 22B at FP16?
Yes. Codestral 22B at FP16 requires 44GB VRAM. Several workstation GPUs (48–96GB) can handle this on a single card.
What quantization is best for Codestral 22B?
Q4_K_M (13GB) offers the best hardware compatibility and still produces high-quality output. Q8_0 (24GB) is better for tasks needing higher accuracy at the cost of needing more VRAM. FP16 (44GB) is only practical on very high-end workstation hardware.