Skip to content
fedi.software

Mistral Medium 3 in fedi-index

The score in every slice, what it is made of, value, local run and history.

fedi-index · General

C 1,341 ±4

Percentile 45 · #162

People prefer / Solves tasks

1,353 / 1,283

Human votes vs benchmarks, Elo-eq

Value

-31

$0.80 per 1M tokens, median of 9 providers · vendor $0.80

On your hardware

Closed weights: cloud only.

Score decomposition

C General 1,341 [1,337–1,345] Percentile 45 · #162 · components: 7
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,370 ±2.40 1,370 ±2 22.8% 2026-10-02 —
LMArena text/hard_prompts 1,365 ±3.04 1,363 ±3 22.6% 2026-10-02 —
Epoch AI gpqa_diamond decay ×0.26 0.60 ±0.03 1,317 ±13 12.2% 2025-05-07 —
LMArena text/instruction_following 1,339 ±3.65 1,355 ±4 11.1% 2026-10-02 —
LMArena text/expert 1,342 ±6.91 1,347 ±6 10.2% 2026-10-02 —
LMArena text/multi_turn 1,383 ±4.19 1,384 ±4 5.5% 2026-10-02 —
LMArena text/longer_query 1,360 ±4.09 1,363 ±4 5.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Coding 1,352 [1,346–1,358] Percentile 46 · #151 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,386 ±4.08 1,368 ±4 90.1% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Math 1,329 [1,319–1,339] Percentile 39 · #156 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,352 ±6.16 1,347 ±6 87.4% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Vision 1,332 [1,327–1,338] Percentile 24 · #86 · components: 3 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena vision/overall 1,157 ±4.21 1,345 ±4 64.4% 2026-10-02 —
LMArena vision/ocr 1,154 ±5.33 1,330 ±5 15.3% 2026-10-02 —
LMArena vision/diagram 1,155 ±8.11 1,335 ±7 13.9% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

B Writing 1,358 [1,351–1,364] Percentile 50 · #134 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,344 ±4.86 1,371 ±5 61.5% 2026-10-02 —
LMArena text/longer_query 1,360 ±4.09 1,363 ±4 32.1% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: 繁體中文 1,343 [1,334–1,353] Percentile 36 · #142 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,374 ±6.65 1,353 ±5 93.4% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: Deutsch 1,369 [1,354–1,384] Percentile 40 · #87 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,377 ±9.54 1,382 ±8 92.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

B Language: English 1,362 [1,356–1,368] Percentile 51 · #135 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,385 ±3.01 1,372 ±3 93.9% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Español 1,350 [1,328–1,371] Percentile 22 · #116 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,363 ±13 1,364 ±12 90.8% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Français 1,342 [1,320–1,364] Percentile 20 · #108 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,367 ±13 1,357 ±12 90.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: 日本語 1,352 [1,338–1,367] Percentile 31 · #89 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 1,317 ±10 1,364 ±8 92.7% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: 한국어 1,349 [1,332–1,366] Percentile 26 · #101 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,303 ±12 1,362 ±9 92.0% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: Polski 1,371 [1,362–1,380] Percentile 28 · #95 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,371 ±5.95 1,383 ±5 93.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

C Language: Русский 1,359 [1,350–1,368] Percentile 42 · #137 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,359 ±5.56 1,370 ±5 93.5% 2026-10-02 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

History

fedi-index by month
The same in a table
Model 2026-10
Mistral Medium 3 1,341 C

Badge

Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.

fedi-index: Mistral Medium 3

Data CC BY 4.0: keep the link to fedi.software.