Skip to content
fedi.software

Gemma 2B IT in fedi-index

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

fedi-index · General

D 1,023 ±10

Percentile 4 · #284

People prefer / Solves tasks

1,024 / 1,018

Human votes vs benchmarks, Elo-eq

Value

—

On your hardware

Closed weights: cloud only.

Score decomposition

D General 1,023 [1,013–1,032] Percentile 4 · #284 · components: 6
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,002 ±5.85 1,002 ±6 29.6% 2026-11-11 —
LMArena text/hard_prompts 989 ±8.82 1,023 ±8 26.7% 2026-11-11 —
LMArena text/instruction_following 970 ±8.16 1,006 ±8 13.5% 2026-11-11 —
Epoch AI eci 76.01 ±6.90 854 ±54 7.2% 2024-02-21 —
LMArena text/multi_turn 945 ±12 986 ±11 5.6% 2026-11-11 —
LMArena text/longer_query 980 ±15 1,009 ±14 4.6% 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

D Coding 1,059 [1,042–1,075] Percentile 5 · #273 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,009 ±11 1,033 ±10 84.8% 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

D Math 1,047 [1,029–1,065] Percentile 4 · #245 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,009 ±11 1,018 ±11 84.1% 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

D Writing 1,026 [1,009–1,042] Percentile 5 · #251 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 986 ±12 1,006 ±12 62.8% 2026-11-11 —
LMArena text/longer_query 980 ±15 1,009 ±14 27.5% 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

D Language: 繁體中文 1,058 [1,040–1,077] Percentile 3 · #213 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 986 ±13 1,045 ±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

D Language: English 1,015 [1,002–1,028] Percentile 5 · #256 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,033 ±6.76 1,000 ±7 92.9% 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

D Language: Русский 1,016 [988–1,044] Percentile 0 · #230 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 936 ±18 990 ±16 87.9% 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

History

fedi-index by month
1,020 1,025 1,030 1,035 2026-10 2026-11 Gemma 2B IT · 2026-10 · 1,031Gemma 2B IT · 2026-11 · 1,023 Gemma 2B IT
The same in a table
Model 2026-102026-11
Gemma 2B IT 1,0311,023 D

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: Gemma 2B IT

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