Skip to content
fedi.software

Hy3 preview in fedi-index

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

fedi-index · General

B 1,379 ±4

Percentile 54 · #136 · provisional

People prefer / Solves tasks

1,391 / —

Human votes vs benchmarks, Elo-eq

Value

+45

$0.2771 per 1M tokens, median of 2 providers

On your hardware

  • 128 GB unified memory #40 · Q2_K · strong compression · ≈16 tok/s

Score decomposition

B General 1,379 [1,375–1,383] Percentile 54 · #136 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,401 ±3.78 1,401 ±4 26.8% 2026-11-11 —
LMArena text/hard_prompts 1,413 ±4.64 1,406 ±4 26.4% 2026-11-11 —
LMArena text/instruction_following 1,384 ±6.40 1,398 ±6 12.2% 2026-11-11 —
LMArena text/expert 1,424 ±11 1,414 ±9 10.4% 2026-11-11 —
LMArena text/longer_query 1,402 ±5.80 1,403 ±5 6.3% 2026-11-11 —
LMArena text/multi_turn 1,410 ±8.96 1,408 ±8 5.5% 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 Coding 1,381 [1,375–1,386] Percentile 55 · #129 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena webdev/overall 1,358 ±8.94 1,409 ±2 46.8% 2026-11-15 —
LMArena text/coding 1,421 ±6.87 1,400 ±6 41.2% 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

B Math 1,377 [1,356–1,398] Percentile 52 · #122 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,424 ±14 1,416 ±13 82.0% 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 Writing 1,367 [1,356–1,379] Percentile 52 · #127 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,340 ±9.72 1,367 ±10 56.5% 2026-11-11 —
LMArena text/longer_query 1,402 ±5.80 1,403 ±5 35.9% 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

C Language: 繁體中文 1,384 [1,363–1,405] Percentile 47 · #117 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,435 ±15 1,402 ±12 90.9% 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: English 1,385 [1,375–1,395] Percentile 55 · #121 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,409 ±5.28 1,398 ±6 93.4% 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 Language: Русский 1,389 [1,370–1,407] Percentile 53 · #109 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,399 ±12 1,406 ±10 91.5% 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
Hy3 preview 1,379 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: Hy3 preview

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