Skip to content
fedi.software

Step 5 Preview in fedi-index

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

fedi-index · General

B 1,419 ±6

Percentile 79 · #63 · provisional

People prefer / Solves tasks

1,434 / —

Human votes vs benchmarks, Elo-eq

Value

+32

$1.42 per 1M tokens, median of 7 providers · vendor $1.40

On your hardware

Closed weights: cloud only.

Score decomposition

B General 1,419 [1,413–1,425] Percentile 79 · #63 · components: 5 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,448 ±5.62 1,448 ±6 31.3% 2026-10-02 —
LMArena text/hard_prompts 1,468 ±6.99 1,456 ±6 30.3% 2026-10-02 —
LMArena text/instruction_following 1,442 ±9.33 1,453 ±9 13.2% 2026-10-02 —
LMArena text/longer_query 1,460 ±8.63 1,456 ±8 6.9% 2026-10-02 —
LMArena text/multi_turn 1,453 ±15 1,447 ±14 4.9% 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

A Coding 1,437 [1,430–1,444] Percentile 88 · #34 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena webdev/overall 1,570 ±6.52 1,470 ±2 51.9% 2026-10-01 high
LMArena text/coding 1,489 ±11 1,460 ±10 37.1% 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

B Agents 1,432 [1,426–1,439] Percentile 51 · #29 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena agent/overall 0.00 ±0.01 1,458 ±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

B Vision 1,434 [1,417–1,451] Percentile 70 · #34 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena vision/overall 1,278 ±12 1,455 ±11 75.7% 2026-10-02 —
LMArena vision/ocr 1,297 ±15 1,468 ±14 15.2% 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 1,431 [1,416–1,446] Percentile 79 · #56 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,422 ±12 1,451 ±13 54.7% 2026-10-02 —
LMArena text/longer_query 1,460 ±8.63 1,456 ±8 36.5% 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

A Language: English 1,432 [1,414–1,450] Percentile 84 · #45 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,461 ±9.43 1,452 ±10 91.7% 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

B Language: Русский 1,425 [1,403–1,447] Percentile 78 · #53 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,447 ±14 1,449 ±12 90.4% 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

History

fedi-index by month
The same in a table
Model 2026-10
Step 5 Preview 1,419 B

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: Step 5 Preview

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