Skip to content
fedi.software

Nemotron 3 Ultra in fedi-index

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

fedi-index · General

B 1,411 ±3

Percentile 74 · #77 · provisional

People prefer / Solves tasks

1,425 / —

Human votes vs benchmarks, Elo-eq

Value

+33

$1 per 1M tokens, median of 14 providers · vendor $1

On your hardware

Does not fit the reference setups.

Score decomposition

B General 1,411 [1,408–1,415] Percentile 74 · #77 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,445 ±3.50 1,445 ±4 26.1% 2026-10-02 —
LMArena text/hard_prompts 1,452 ±3.94 1,442 ±4 26.0% 2026-10-02 —
LMArena text/instruction_following 1,418 ±5.05 1,430 ±5 12.5% 2026-10-02 —
LMArena text/expert 1,472 ±8.65 1,453 ±7 11.3% 2026-10-02 —
LMArena text/longer_query 1,438 ±4.57 1,436 ±4 6.4% 2026-10-02 —
LMArena text/multi_turn 1,412 ±7.44 1,410 ±7 5.7% 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 Coding 1,417 [1,408–1,426] Percentile 76 · #68 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,468 ±5.55 1,442 ±5 89.6% 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 Math 1,405 [1,386–1,424] Percentile 75 · #65 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,455 ±12 1,446 ±12 83.1% 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 Writing 1,414 [1,405–1,422] Percentile 72 · #77 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,397 ±6.87 1,426 ±7 59.6% 2026-10-02 —
LMArena text/longer_query 1,438 ±4.57 1,436 ±4 33.6% 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: 繁體中文 1,433 [1,417–1,449] Percentile 79 · #47 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,498 ±11 1,452 ±9 92.2% 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

A Language: English 1,430 [1,422–1,439] Percentile 83 · #48 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,455 ±4.42 1,446 ±5 93.6% 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: Español 1,424 [1,401–1,448] Percentile 71 · #44 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,455 ±14 1,449 ±13 89.9% 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

B Language: Français 1,428 [1,404–1,451] Percentile 71 · #40 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,472 ±15 1,453 ±13 90.0% 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 1,404 [1,390–1,419] Percentile 63 · #88 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,415 ±8.89 1,421 ±8 92.6% 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
Nemotron 3 Ultra 1,411 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: Nemotron 3 Ultra

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