Skip to content
fedi.software

Qwen3.5-27B in fedi-index

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

fedi-index · General

B 1,386 ±2

Percentile 60 · #118 · provisional

People prefer / Solves tasks

1,397 / —

Human votes vs benchmarks, Elo-eq

Value

+14

$0.7838 per 1M tokens, median of 18 providers · vendor $0.825

On your hardware

  • GPU 12 GB #18 · Q4_K_M · ≈4 tok/s
  • GPU 24 GB #5 · Q5_K_S · ≈38 tok/s
  • Mac 64 GB #9 · Q8_0 · ≈13 tok/s
  • 128 GB unified memory #20 · Q4_K_M · ≈10 tok/s
  • CPU, 32 GB RAM #10 · Q4_K_M · ≈2 tok/s

Score decomposition

B General 1,386 [1,384–1,388] Percentile 60 · #118 · components: 6 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,409 ±2.18 1,409 ±2 25.8% 2026-10-02 —
LMArena text/hard_prompts 1,414 ±2.66 1,408 ±2 25.7% 2026-10-02 —
LMArena text/instruction_following 1,393 ±3.38 1,406 ±3 12.6% 2026-10-02 —
LMArena text/expert 1,428 ±5.92 1,417 ±5 11.9% 2026-10-02 —
LMArena text/longer_query 1,413 ±3.26 1,412 ±3 6.3% 2026-10-02 —
LMArena text/multi_turn 1,410 ±4.51 1,408 ±4 6.1% 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 Coding 1,387 [1,384–1,390] Percentile 61 · #109 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena webdev/overall 1,357 ±4.09 1,409 ±1 46.1% 2026-10-01 —
LMArena text/coding 1,427 ±3.62 1,405 ±3 44.3% 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

B Math 1,392 [1,380–1,404] Percentile 65 · #91 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,429 ±7.15 1,421 ±7 86.9% 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 Vision 1,406 [1,403–1,410] Percentile 50 · #57 · components: 3 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena vision/overall 1,240 ±2.85 1,421 ±3 62.9% 2026-10-02 —
LMArena vision/ocr 1,245 ±2.81 1,418 ±3 15.7% 2026-10-02 —
LMArena vision/diagram 1,248 ±4.03 1,417 ±4 15.3% 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

B Writing 1,386 [1,379–1,392] Percentile 58 · #113 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,362 ±4.78 1,390 ±5 61.2% 2026-10-02 —
LMArena text/longer_query 1,413 ±3.26 1,412 ±3 32.5% 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

B Language: 繁體中文 1,421 [1,411–1,430] Percentile 73 · #60 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,478 ±6.73 1,436 ±5 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

C Language: Deutsch 1,378 [1,357–1,399] Percentile 43 · #83 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,393 ±14 1,396 ±12 90.7% 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 Language: English 1,401 [1,396–1,407] Percentile 63 · #101 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,425 ±2.98 1,414 ±3 93.9% 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

C Language: Español 1,389 [1,372–1,405] Percentile 40 · #90 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,407 ±9.91 1,405 ±9 92.1% 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: Français 1,379 [1,361–1,397] Percentile 34 · #90 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,409 ±11 1,395 ±10 91.8% 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

C Language: 한국어 1,386 [1,366–1,406] Percentile 44 · #77 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/korean 1,358 ±15 1,403 ±11 91.1% 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: Polski 1,371 [1,351–1,390] Percentile 27 · #96 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/polish 1,375 ±13 1,386 ±11 91.4% 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

B Language: Русский 1,385 [1,376–1,394] Percentile 52 · #113 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,390 ±5.28 1,398 ±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

History

fedi-index by month
The same in a table
Model 2026-10
Qwen3.5-27B 1,386 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: Qwen3.5-27B

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