Skip to content
fedi.software

Grok 4.20 Beta 1 in fedi-index

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

fedi-index · General

B 1,407 ±2

Percentile 70 · #91 · provisional

People prefer / Solves tasks

1,419 / —

Human votes vs benchmarks, Elo-eq

Value

—

On your hardware

Closed weights: cloud only.

Score decomposition

B General 1,407 [1,404–1,409] Percentile 70 · #91 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,444 ±2.32 1,444 ±2 25.8% 2026-11-11 —
LMArena text/hard_prompts 1,441 ±2.78 1,432 ±2 25.7% 2026-11-11 —
LMArena text/instruction_following 1,415 ±3.56 1,428 ±3 12.6% 2026-11-11 —
LMArena text/expert 1,428 ±6.29 1,417 ±5 11.8% 2026-11-11 —
LMArena text/longer_query 1,426 ±3.45 1,425 ±3 6.3% 2026-11-11 —
LMArena text/multi_turn 1,451 ±4.77 1,445 ±4 6.1% 2026-11-11 —

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 Coding 1,397 [1,391–1,403] Percentile 63 · #107 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,447 ±3.83 1,423 ±3 88.4% 2026-11-11 —

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 Math 1,395 [1,383–1,407] Percentile 63 · #95 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,433 ±7.35 1,425 ±7 86.8% 2026-11-11 —

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 Search 1,418 [1,412–1,424] Percentile 37 · #23 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena search/overall 1,191 ±3.08 1,432 ±3 93.9% 2026-11-09 —

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 1,437 [1,431–1,444] Percentile 81 · #51 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,439 ±4.94 1,468 ±5 61.1% 2026-11-11 —
LMArena text/longer_query 1,426 ±3.45 1,425 ±3 32.5% 2026-11-11 —

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 1,419 [1,408–1,430] Percentile 69 · #68 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,476 ±7.51 1,435 ±6 93.2% 2026-11-11 —

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: Deutsch 1,434 [1,412–1,455] Percentile 81 · #29 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,464 ±14 1,457 ±12 90.7% 2026-11-11 —

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 1,427 [1,421–1,433] Percentile 79 · #57 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,451 ±3.09 1,442 ±3 93.9% 2026-11-11 —

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 1,415 [1,399–1,432] Percentile 58 · #64 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,438 ±9.94 1,434 ±9 92.1% 2026-11-11 —

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 1,418 [1,400–1,436] Percentile 57 · #61 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,455 ±11 1,438 ±10 91.7% 2026-11-11 —

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 1,428 [1,408–1,447] Percentile 74 · #37 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,418 ±14 1,449 ±11 91.3% 2026-11-11 —

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 1,416 [1,398–1,434] Percentile 55 · #59 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,433 ±12 1,436 ±10 91.6% 2026-11-11 —

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 1,436 [1,426–1,444] Percentile 80 · #48 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,450 ±5.43 1,452 ±5 93.5% 2026-11-11 —

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 2026-11
Grok 4.20 Beta 1 1,407 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: Grok 4.20 Beta 1

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