Skip to content
fedi.software

NVIDIA-Nemotron-3.5-Lightning-30B-A3B (NVFP4) in fedi-index

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

fedi-index · General

C 1,322 ±3

Percentile 39 · #181 · provisional

People prefer / Solves tasks

1,329 / —

Human votes vs benchmarks, Elo-eq

Value

+18◆

$0.1025 per 1M tokens, median of 3 providers

On your hardware

Closed weights: cloud only.

Score decomposition

C General 1,322 [1,320–1,325] Percentile 39 · #181 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,329 ±2.70 1,330 ±3 25.8% 2026-11-11 nvfp4
LMArena text/hard_prompts 1,340 ±3.14 1,341 ±3 25.7% 2026-11-11 nvfp4
LMArena text/instruction_following 1,318 ±3.90 1,336 ±4 12.6% 2026-11-11 nvfp4
LMArena text/expert 1,360 ±6.28 1,362 ±5 11.9% 2026-11-11 nvfp4
LMArena text/longer_query 1,319 ±3.71 1,325 ±4 6.3% 2026-11-11 nvfp4
LMArena text/multi_turn 1,324 ±5.74 1,330 ±5 5.9% 2026-11-11 nvfp4

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 Coding 1,346 [1,339–1,352] Percentile 41 · #171 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,381 ±4.30 1,365 ±4 88.2% 2026-11-11 nvfp4

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 Math 1,338 [1,323–1,353] Percentile 40 · #154 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,366 ±9.41 1,361 ±9 85.4% 2026-11-11 nvfp4

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 Writing 1,292 [1,286–1,299] Percentile 37 · #167 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,259 ±4.97 1,285 ±5 61.2% 2026-11-11 nvfp4
LMArena text/longer_query 1,319 ±3.71 1,325 ±4 32.4% 2026-11-11 nvfp4

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 1,336 [1,326–1,347] Percentile 35 · #142 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,365 ±7.32 1,346 ±6 93.3% 2026-11-11 nvfp4

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: Deutsch 1,308 [1,284–1,332] Percentile 26 · #108 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,305 ±16 1,320 ±14 89.6% 2026-11-11 nvfp4

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: English 1,343 [1,336–1,350] Percentile 44 · #150 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,367 ±3.55 1,353 ±4 93.8% 2026-11-11 nvfp4

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: Español 1,341 [1,323–1,360] Percentile 15 · #127 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,352 ±11 1,354 ±10 91.6% 2026-11-11 nvfp4

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: Français 1,353 [1,335–1,371] Percentile 21 · #113 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,378 ±11 1,367 ±10 91.7% 2026-11-11 nvfp4

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: 日本語 1,297 [1,275–1,319] Percentile 20 · #102 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 1,240 ±17 1,307 ±12 90.4% 2026-11-11 nvfp4

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: 한국어 1,311 [1,290–1,333] Percentile 20 · #110 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,252 ±16 1,323 ±12 90.7% 2026-11-11 nvfp4

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: Polski 1,288 [1,264–1,313] Percentile 2 · #128 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,271 ±17 1,298 ±14 89.5% 2026-11-11 nvfp4

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 1,273 [1,263–1,283] Percentile 25 · #173 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,257 ±6.19 1,278 ±6 93.4% 2026-11-11 nvfp4

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
NVIDIA-Nemotron-3.5-Lightning-30B-A3B (NVFP4) 1,322 C

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: NVIDIA-Nemotron-3.5-Lightning-30B-A3B (NVFP4)

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