Skip to content
fedi.software

Qwen3 30B A3B in fedi-index

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

fedi-index · General

C 1,309 ±9

Percentile 36 · #187

People prefer / Solves tasks

1,309 / 1,310

Human votes vs benchmarks, Elo-eq

Value

-29

$0.1975 per 1M tokens, median of 10 providers

On your hardware

  • GPU 12 GB #14 · Q8_0 · ≈16 tok/s
  • GPU 24 GB #18 · Q5_K_S · ≈147 tok/s
  • Mac 64 GB #16 · Q8_0 · ≈45 tok/s
  • 128 GB unified memory #27 · Q8_0 · ≈33 tok/s
  • CPU, 32 GB RAM #9 · Q4_K_M · ≈6 tok/s

Score decomposition

C General 1,309 [1,300–1,318] Percentile 36 · #187 · components: 8
Source Component Value Elo-eq Weight Measured Variant
Epoch AI eci 136 ±2.28 1,329 ±18 24.5% 2025-04-29 —
LMArena text/overall 1,317 ±2.41 1,317 ±2 19.4% 2026-10-02 —
LMArena text/hard_prompts 1,314 ±3.24 1,317 ±3 19.2% 2026-10-02 —
LMArena text/instruction_following 1,284 ±3.99 1,303 ±4 9.4% 2026-10-02 —
LMArena text/expert 1,313 ±8.30 1,324 ±7 8.3% 2026-10-02 —
LMArena text/multi_turn 1,307 ±4.64 1,314 ±4 4.6% 2026-10-02 —
LMArena text/longer_query 1,312 ±4.53 1,318 ±4 4.6% 2026-10-02 —
Epoch AI gpqa_diamond decay ×0.26 0.62 ±0.03 1,327 ±14 1.4% 2026-08-30 —

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 1,313 [1,306–1,320] Percentile 34 · #185 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,337 ±4.50 1,325 ±4 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

C Math 1,330 [1,319–1,341] Percentile 39 · #155 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,354 ±6.95 1,349 ±7 87.0% 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

C Writing 1,298 [1,291–1,305] Percentile 36 · #171 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,271 ±5.30 1,297 ±5 61.4% 2026-10-02 —
LMArena text/longer_query 1,312 ±4.53 1,318 ±4 32.1% 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

C Language: 繁體中文 1,323 [1,312–1,334] Percentile 31 · #154 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,347 ±7.64 1,332 ±6 93.2% 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

C Language: Deutsch 1,313 [1,296–1,330] Percentile 25 · #108 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,308 ±11 1,323 ±9 92.0% 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

C Language: English 1,312 [1,305–1,318] Percentile 38 · #171 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,334 ±3.12 1,319 ±3 93.8% 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

D Language: Español 1,310 [1,287–1,333] Percentile 12 · #131 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,317 ±14 1,322 ±13 90.1% 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

D Language: Français 1,326 [1,302–1,351] Percentile 15 · #115 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,350 ±15 1,341 ±14 89.4% 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

D Language: 日本語 1,309 [1,294–1,325] Percentile 22 · #100 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 1,255 ±11 1,318 ±8 92.4% 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

D Language: 한국어 1,319 [1,300–1,338] Percentile 22 · #107 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,262 ±14 1,330 ±10 91.5% 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

D Language: Polski 1,314 [1,303–1,325] Percentile 6 · #123 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,299 ±7.01 1,322 ±6 93.3% 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

C Language: Русский 1,302 [1,291–1,312] Percentile 30 · #164 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,291 ±6.11 1,308 ±6 93.4% 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
Qwen3 30B A3B 1,309 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: Qwen3 30B A3B

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