Pass IndexThe State of AISign in

Magistral Small 2506

Magistral Small 2506 is a model by Mistral AI for reasoning, it stands 197th of 205 on ARC-AGI-1, and one seller in the catalogue publishes a price for it.

text → text · made by Mistral AI

$0.5$1.5per Mtok in / out
IO.NET

Sold by 1 way

SellerLaneRate
IO.NET aggregatorstandard$0.5per Mtok in$1.5per Mtok out$0.25per Mtok cached$1per Mtok cache writemodels.dev · read 2026-09-16

Measured 9 standings

PlaceBoardMetricScore
197thof 205ARC-AGI-1 (Public Eval)Score (%)8.62
199thof 216ARC-AGI-2 (Public Eval)Score (%)0
213thof 313GPQA Diamondmean_score56.061
214thof 290OTIS Mock AIME 2024-2025mean_score30
219thof 227Arc agi 2 — Epoch AIScore0
222ndof 230ARC-AGI-1 (Semi-Private)Score (%)5
225thof 232ARC-AGI-2 (Semi-Private)Score (%)0
237thof 244Arc agi — Epoch AIScore5
509thof 656Epoch capabilities index — Epoch AIECI Score133.21

About

Magistral Small 2506 — a text model from Mistral AI, sold by one company from $0.5 in and $1.5 out per million tokens, placed 213th of 313 on GPQA Diamond.

It takes text and returns text, with a context window of 128,000 tokens. It was published in June 2025, trained on material up to January 2025. The catalogue files it under reasoning. Only IO.NET sells it, at $0.5 in and $1.5 out per million tokens. It has been measured on 9 boards, and stands best at 213th of 313 on GPQA Diamond.

Every current figure

Maker
Mistral AI
Register
model
Takes
text
Returns
text
Context
128,000 tokens
Longest answer
32,768 tokens
Published
June 2025
Knowledge to
January 2025
Licence
not read
Sellers
1
Maker's own price
not read
Price
$0.5 in and $1.5 out per million tokens — IO.NET
Boards
9
Best place
213th of 313 — GPQA Diamond