Skip to content
fedi.software

Claude Sonnet 5 in fedi-index

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

fedi-index · General

A 1,443 ±6

Percentile 89 · #34

People prefer / Solves tasks

1,438 / 1,452

Human votes vs benchmarks, Elo-eq

Value

+32

$4 per 1M tokens, median of 31 providers · vendor $4

On your hardware

Closed weights: cloud only.

Score decomposition

A General 1,443 [1,437–1,450] Percentile 89 · #34 · components: 9
Source Component Value Elo-eq Weight Measured Variant
Epoch AI eci 156 ±1.35 1,487 ±11 30.1% 2026-06-30 —
LMArena text/overall 1,443 ±2.14 1,443 ±2 17.0% 2026-10-02 high
LMArena text/hard_prompts 1,462 ±2.48 1,451 ±2 16.9% 2026-10-02 high
LMArena text/instruction_following 1,453 ±2.96 1,463 ±3 8.4% 2026-10-02 high
LMArena text/expert 1,491 ±4.52 1,469 ±4 8.2% 2026-10-02 high
LMArena text/longer_query 1,464 ±2.81 1,460 ±3 4.2% 2026-10-02 high
LMArena text/multi_turn 1,454 ±3.84 1,448 ±4 4.1% 2026-10-02 high
Epoch AI simpleqa_verified outlier decay ×0.61 0.34 ±0.01 1,403 ±5 2.1% 2026-08-27 max
Epoch AI gpqa_diamond decay ×0.26 0.91 ±0.02 1,461 ±8 1.5% 2026-07-01 xhigh

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 1,433 [1,431–1,436] Percentile 84 · #44 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena webdev/overall 1,539 ±3.16 1,461 ±1 45.9% 2026-10-01 high
LMArena text/coding 1,483 ±3.23 1,455 ±3 44.5% 2026-10-02 high

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 1,449 [1,442–1,456] Percentile 74 · #16 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena agent/overall 0.04 ±0.01 1,476 ±4 90.0% 2026-10-02 high

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 Math 1,438 [1,430–1,446] Percentile 88 · #31 · components: 3
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,467 ±6.72 1,458 ±6 51.3% 2026-10-02 high
Epoch AI frontiermath_t1_3 decay ×0.86 0.66 ±0.03 1,463 ±6 30.0% 2026-06-30 max
Epoch AI frontiermath_t4 decay ×0.86 0.29 ±0.07 1,439 ±12 11.0% 2026-06-30 max

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 1,440 [1,435–1,444] Percentile 74 · #30 · components: 3 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena vision/overall 1,275 ±3.57 1,452 ±3 63.2% 2026-10-02 high
LMArena vision/ocr 1,289 ±3.71 1,461 ±4 15.6% 2026-10-02 high
LMArena vision/diagram 1,298 ±5.52 1,461 ±5 14.9% 2026-10-02 high

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 Writing 1,435 [1,430–1,440] Percentile 81 · #52 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,416 ±3.71 1,445 ±4 61.9% 2026-10-02 high
LMArena text/longer_query 1,464 ±2.81 1,460 ±3 31.9% 2026-10-02 high

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: 繁體中文 1,420 [1,412–1,428] Percentile 72 · #62 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,476 ±5.42 1,434 ±4 93.7% 2026-10-02 high

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: Deutsch 1,420 [1,402–1,438] Percentile 69 · #45 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,444 ±12 1,440 ±10 91.7% 2026-10-02 high

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 1,428 [1,422–1,433] Percentile 80 · #54 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,451 ±2.76 1,442 ±3 93.9% 2026-10-02 high

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: Español 1,414 [1,399–1,429] Percentile 58 · #63 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,435 ±8.90 1,431 ±8 92.5% 2026-10-02 high

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: Français 1,422 [1,408–1,436] Percentile 67 · #45 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,457 ±8.66 1,440 ±8 92.6% 2026-10-02 high

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: 日本語 1,422 [1,405–1,440] Percentile 73 · #35 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 1,420 ±13 1,442 ±10 91.9% 2026-10-02 high

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: 한국어 1,425 [1,410–1,441] Percentile 73 · #38 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,411 ±11 1,444 ±8 92.4% 2026-10-02 high

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: Polski 1,407 [1,389–1,425] Percentile 50 · #66 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,422 ±12 1,426 ±10 91.7% 2026-10-02 high

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: Русский 1,437 [1,429–1,444] Percentile 83 · #41 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,451 ±4.48 1,452 ±4 93.7% 2026-10-02 high

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
Claude Sonnet 5 1,443 A

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: Claude Sonnet 5

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