How much VRAM does Devstral 2 123B need?

Devstral 2 123B requires approximately 0.0 GB VRAM at Q4_K_M quantization with 1K context and 10% overhead. At Q8_0 it needs ~0.0 GB, and at FP16 ~0.0 GB. Use the PCPARTGUIDE VRAM Calculator for exact requirements at any context length and overhead setting.

Can I run Devstral 2 123B on a consumer GPU?

Devstral 2 123B is a 123B parameter model and requires datacenter or multi-GPU setups. At Q4_K_M, it needs approximately 0.0 GB VRAM, which exceeds consumer GPU limits. It can run on H100 (80 GB) or A100 (80 GB) datacenter GPUs, or across multiple consumer GPUs with tensor parallelism.

What is Devstral 2 123B's context window?

Devstral 2 123B supports a maximum context window of 262,144 tokens. The default context is 4,096 tokens. Larger contexts require proportionally more VRAM for the KV cache — use the VRAM Calculator to estimate exact requirements at your target context length.

What quantization level should I use for Devstral 2 123B?

For Devstral 2 123B, start with Q4_K_M (~0.0 GB) if VRAM-limited. Q8_0 offers better quality at ~2× the VRAM. If you have enough VRAM, Q6_K (~1.5× Q4 VRAM) provides a good balance. For models this large, Q3_K_M may be necessary to fit on consumer hardware.

How does Devstral 2 123B compare to similar models?

Devstral 2 123B is a 123.0B Dense model from the Mistral family. Browse other Mistral models on the PCPARTGUIDE directory for comparable alternatives with different parameter counts and VRAM requirements. Use the Model Transparency scores to assess documentation quality and reproducibility.

MistralDenseMIT

Devstral 2 123B

Devstral 2 123B is a dense transformer language model from the Mistral family, containing 123B parameters across 96 layers. It supports up to 262K tokens of context with a hidden dimension of 10240 and 16 KV heads for efficient grouped-quer…

123.0B

Parameters

256K

Max Context

Dense

Architecture

—

Released

Text

Modality

About Devstral 2 123B

CodeAgentic

Technical Specifications

Total Parameters123.0B

ArchitectureDense

Attention TypeGQA (Grouped Query Attention)

Hidden Dimensiond = 10,240

Transformer Layers96

Attention Heads80

KV Headsn_kv = 16

Head Dimensiond_head = 128

Activation FunctionSwiGLU

NormalizationRMSNorm

Position EmbeddingRoPE

System Requirements

Estimated VRAM at 10% overhead for different quantization methods and context sizes.

Quantization	1K ctx	195K ctx	256K ctx
Q4_K_M0.50 B/W ~97% of FP16	64.33Datacenter GPU	210.1Cluster / Multi-GPU	255.6Cluster / Multi-GPU
Q8_01.00 B/W ~100% of FP16	127.9Cluster / Multi-GPU	273.6Cluster / Multi-GPU	319.2Cluster / Multi-GPU
F162.00 B/W Reference	255.1Cluster / Multi-GPU	400.8Cluster / Multi-GPU	446.3Cluster / Multi-GPU

Params

675.0B

Layers

Context

256K

41.0B active

Find the right GPU for Devstral 2 123B

Use the interactive VRAM Calculator to see exactly how much memory you need at any quantization level, context length, and overhead setting.

Open VRAM Calculator Best GPU Guide VRAM Guide