AI Index

Devstral Small 1.0

Language Mistral AI Partly verified Retired Shut down Nov 30, 2025 Released May 21, 2025
  • In Text Out Text
  • Tool use
  • Structured output
  • Open weights

Prices

1 offering · prices as listed
Mistral AI API Official Retired Global · USD devstral-small-2505
Conditions/ 1M tokensInputOutput
Standard $0.10$0.30

Prices from Devstral | Mistral AI (checked Oct 30, 2026)

Standard: the price a request with no conditions pays (the headline price elsewhere on the site).

Details

Reasoning
—
Knowledge cutoff
May 2025
Licence
Apache 2.0
Family
devstral-small
Model ID

Not stated on Mistral AI’s pages: max input, reasoning efforts, reasoning can be off, max thinking budget, built-in tools, aliases.

Run locally

Weights: mistralai/Devstral-Small-2505
  • Parameters 23.6B
  • Dense
  • Native precision BF16
  • Layers 40
  • KV cache per token 160 KB
  • Config positions 128K (config.json max_position_embeddings)
  • API context 125K (maker’s API)
My hardware GeForce RTX 4090 24 GB · 1008 GB/s · 64 GB DDR5 dual channel · 89.6 GB/s
On your hardware · GeForce RTX 4090
Runs well Q5_K_M (automatic) 8K context 28–42 tok/sOne GPU
WeightsKV cacheOverheadUsable
Needs
≈ 17.6 GB (weights 15.6 + KV cache 1.3 + overhead 0.8)
Usable
21.6 GB (90% of its memory)
Generation speed
28–42 tok/s
Longest context
33K (longer doesn’t fit)
First token
about 0.76–1.14 s for a 2K prompt
PrecisionWeightsNeeds at 8KOn your hardwareLongest context
BF16 (native)maker’s file87.8 GB≈ 93.5Doesn’t fit—
Q8_023.3 GB≈ 25.7Needs system memory 7–10 tok/s—
FP822.0 GB≈ 24.3Needs system memory 9–14 tok/s—
Q6_K18.0 GB≈ 20.2Runs well 24–37 tok/s17K
Q5_K_M15.6 GB≈ 17.6Runs well 28–42 tok/s33K
Q5_015.1 GB≈ 17.1Runs well 29–43 tok/s36K
Q4_K_M13.3 GB≈ 15.3Runs well 32–48 tok/s48K
Q4_012.3 GB≈ 14.2Runs well 35–52 tok/s55K
IQ4_XS11.7 GB≈ 13.5Runs well 36–55 tok/s60K
Q3_K_Mquality drops10.7 GB≈ 12.5Runs well 39–59 tok/s66K
IQ3_XXSquality drops8.4 GB≈ 10.1Runs well 49–73 tok/s81K
Q2_Kquality drops8.2 GB≈ 10.0Runs well 49–74 tok/s82K
IQ2_XXSquality drops5.7 GB≈ 7.4Runs well 68–102 tok/s99K

Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.

Compared with other hardware

Each at the highest precision it runs. Click one for details.

Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 4090; unified memory Apple M5 Pro (16-core GPU) · 24 GB.

Run commands

The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.

Memory to fine-tune
Full≈ 353.9 GBDoesn’t fit your hardware
LoRA≈ 46.8 GBDoesn’t fit your hardware
QLoRA≈ 15.3 GBFits your hardware

Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.

Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated

Cited scores

From third-party leaderboards, not measured by us

This model isn’t on any leaderboard we cite yet.

Changes

All changes →
  1. New language model

Sources

Mistral AI’s official policies: data use, privacy, terms →

Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data

Mistral AI · France · mistral.ai

Snapshots, the same family and other models from this provider