Devstral Small 2507 — Hardware Requirements & GPU Compatibility
Chat
Devstral Small 2507 is Mistral AI's agentic coding model, developed with All Hands AI and fine-tuned from the 24-billion-parameter Mistral Small 3.1 with its vision encoder removed to keep it text-only. It is built to explore codebases, edit multiple files, and drive software-engineering agents, using Mistral's function-calling format and a Tekken tokenizer with a 131K-token vocabulary. It supports a 128K token context window and is released under the Apache 2.0 license. At 24 billion parameters, Devstral is light enough to run on a single RTX 4090 or a Mac with around 32 GB of unified memory once quantized to 4-bit, making it practical for local coding-agent setups.
72.3K downloads
368 likesSpecifications
- Publisher
- Mistral AI
- Family
- Mistral
- Parameters
- 23.6B
- Architecture
- MistralForCausalLM
- Context Length
- 131,072 tokens
- Vocabulary Size
- 131,072
- Release Date
- 2025-07-04
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Devstral Small 2507 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 10.7 GB | 37.2 GB | 10.02 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 11.0 GB | 37.5 GB | 10.31 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 12.2 GB | 38.6 GB | 11.49 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 12.5 GB | 38.9 GB | 11.79 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 14.9 GB | 41.3 GB | 14.14 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 17.5 GB | 43.9 GB | 16.80 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 20.2 GB | 46.6 GB | 19.45 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 24.3 GB | 50.7 GB | 23.57 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Devstral Small 2507?
Q4_K_M · 14.9 GBDevstral Small 2507 (Q4_K_M) requires 14.9 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 20+ GB is recommended. Using the full 131K context window can add up to 26.4 GB, bringing total usage to 41.3 GB. 26 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 5080.
Runs great
— Plenty of headroom
Decent
— Enough VRAM, may be tightWhich Devices Can Run Devstral Small 2507?
Q4_K_M · 14.9 GB47 devices with unified memory can run Devstral Small 2507, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Mini M4 (16 GB).
Runs great
— Plenty of headroomWhere to Download Devstral Small 2507
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Benchmarks
Benchmark details →Related Models
Frequently Asked Questions
- How much VRAM does Devstral Small 2507 need?
Devstral Small 2507 requires 14.9 GB of VRAM at Q4_K_M, or 47.9 GB at BF16. Full 131K context adds up to 26.4 GB (41.3 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 23.6B × 4.8 bits ÷ 8 = 14.1 GB
KV Cache + Overhead ≈ 0.8 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 27.2 GB (at full 131K context)
VRAM usage by quantization
Q4_K_M14.9 GBQ4_K_M + full context41.3 GB- Can NVIDIA GeForce RTX 4090 run Devstral Small 2507?
Yes, at Q6_K (20.2 GB) or lower. Higher quantizations like Q8_0 (24.3 GB) exceed the NVIDIA GeForce RTX 4090's 24 GB.
- What's the best quantization for Devstral Small 2507?
For Devstral Small 2507, Q4_K_M (14.9 GB) offers the best balance of quality and VRAM usage. Q4_K_L (15.2 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 7.2 GB.
VRAM requirement by quantization
IQ2_XXS7.2 GBIQ3_S10.7 GBQ3_K_L12.8 GBQ4_K_M ★14.9 GBQ5_015.4 GBBF1647.9 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Devstral Small 2507 on a Mac?
Devstral Small 2507 requires at least 7.2 GB at IQ2_XXS, 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 Devstral Small 2507 locally?
Yes — Devstral Small 2507 can run locally on consumer hardware. At Q4_K_M quantization it needs 14.9 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Devstral Small 2507?
At Q4_K_M, Devstral Small 2507 can reach ~323 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~44 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 14.9 × 0.65 = ~350 tok/s
Estimated speed at Q4_K_M (14.9 GB)
~350 tok/s~44 tok/s~350 tok/s~323 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Devstral Small 2507?
At Q4_K_M, the download is about 14.14 GB. The full-precision BF16 version is 47.14 GB. The smallest option (IQ2_XXS) is 6.48 GB.
- Which GPUs can run Devstral Small 2507?
26 consumer GPUs can run Devstral Small 2507 at Q4_K_M (14.9 GB). Top options include AMD Radeon RX 7900 XTX, NVIDIA GeForce RTX 3090, NVIDIA GeForce RTX 3090 Ti, AMD Radeon RX 6800. 7 GPUs have plenty of headroom for comfortable inference.
- Which devices can run Devstral Small 2507?
49 devices with unified memory can run Devstral Small 2507 at Q4_K_M (14.9 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Asus ROG Flow Z13 (2025, Ryzen AI Max+ 395, 128 GB), Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.