Qwen3 VL 235B A22B Thinking in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
B 1,374 ±4
Percentile 55 · #133 · provisional
People prefer / Solves tasks
1,386 / —
Human votes vs benchmarks, Elo-eq
Value
-17
$1.72 per 1M tokens, median of 10 providers
On your hardware
- GPU 12 GB #29 · UD-TQ1_0 · strong compression · ≈10 tok/s
- GPU 24 GB #32 · UD-IQ1_M · strong compression · ≈9 tok/s
- 128 GB unified memory #25 · UD-Q3_K_XL · ≈12 tok/s
Score decomposition
B General 1,374 [1,370–1,378] Percentile 55 · #133 · components: 6 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/overall
|
1,401 ±3.49 | 1,401 ±4 | 27.5% | 2026-10-02 | — |
| LMArena |
text/hard_prompts
|
1,406 ±4.69 | 1,401 ±4 | 26.8% | 2026-10-02 | — |
| LMArena |
text/instruction_following
|
1,374 ±6.29 | 1,389 ±6 | 12.5% | 2026-10-02 | — |
| LMArena |
text/expert
|
1,427 ±15 | 1,416 ±12 | 8.7% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,391 ±6.95 | 1,392 ±6 | 6.1% | 2026-10-02 | — |
| LMArena |
text/multi_turn
|
1,387 ±8.47 | 1,387 ±8 | 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 1,381 [1,370–1,392] Percentile 58 · #118 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/coding
|
1,425 ±7.30 | 1,403 ±6 | 89.0% | 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 Math 1,366 [1,343–1,388] Percentile 52 · #124 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,411 ±15 | 1,404 ±14 | 81.1% | 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 Vision 1,376 [1,368–1,384] Percentile 39 · #69 · components: 3 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
vision/overall
|
1,206 ±6.25 | 1,390 ±6 | 66.5% | 2026-10-02 | — |
| LMArena |
vision/ocr
|
1,212 ±8.13 | 1,386 ±8 | 14.9% | 2026-10-02 | — |
| LMArena |
vision/diagram
|
1,214 ±14 | 1,386 ±12 | 11.5% | 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 Writing 1,367 [1,355–1,379] Percentile 54 · #125 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,347 ±9.32 | 1,374 ±10 | 58.1% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,391 ±6.95 | 1,392 ±6 | 34.3% | 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 Language: 繁體中文 1,388 [1,368–1,408] Percentile 52 · #108 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,441 ±14 | 1,406 ±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
B Language: English 1,395 [1,386–1,405] Percentile 62 · #105 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,420 ±5.05 | 1,409 ±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: Español 1,379 [1,354–1,405] Percentile 32 · #101 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,402 ±16 | 1,401 ±15 | 89.1% | 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: Polski 1,363 [1,342–1,384] Percentile 22 · #102 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,367 ±14 | 1,379 ±12 | 90.8% | 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: Русский 1,364 [1,346–1,383] Percentile 44 · #131 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,369 ±11 | 1,379 ±10 | 91.6% | 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
The same in a table
| Model | 2026-10 |
|---|---|
| Qwen3 VL 235B A22B Thinking | 1,374 B |
Badge
Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.
Data CC BY 4.0: keep the link to fedi.software.