Qwen3 VL 235B A22B Instruct in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
B 1,394 ±4
Percentile 65 · #104 · provisional
People prefer / Solves tasks
1,407 / —
Human votes vs benchmarks, Elo-eq
Value
+27
$0.6325 per 1M tokens, median of 19 providers · vendor $1.22
On your hardware
- GPU 12 GB #27 · UD-TQ1_0 · strong compression · ≈10 tok/s
- GPU 24 GB #30 · UD-IQ1_M · strong compression · ≈9 tok/s
- 128 GB unified memory #18 · UD-Q3_K_XL · ≈12 tok/s
Score decomposition
B General 1,394 [1,390–1,398] Percentile 65 · #104 · components: 6 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/overall
|
1,420 ±3.33 | 1,420 ±3 | 27.0% | 2026-10-02 | — |
| LMArena |
text/hard_prompts
|
1,428 ±4.40 | 1,420 ±4 | 26.5% | 2026-10-02 | — |
| LMArena |
text/instruction_following
|
1,406 ±5.79 | 1,419 ±6 | 12.5% | 2026-10-02 | — |
| LMArena |
text/expert
|
1,440 ±12 | 1,427 ±10 | 9.7% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,420 ±6.56 | 1,419 ±6 | 6.1% | 2026-10-02 | — |
| LMArena |
text/multi_turn
|
1,428 ±7.35 | 1,424 ±7 | 5.9% | 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 Coding 1,392 [1,382–1,403] Percentile 64 · #101 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/coding
|
1,438 ±6.64 | 1,415 ±6 | 89.2% | 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
B Math 1,383 [1,364–1,401] Percentile 59 · #104 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,426 ±12 | 1,418 ±11 | 83.8% | 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
B Vision 1,412 [1,408–1,417] Percentile 53 · #54 · components: 3 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
vision/overall
|
1,247 ±3.54 | 1,427 ±3 | 63.4% | 2026-10-02 | — |
| LMArena |
vision/ocr
|
1,253 ±4.00 | 1,426 ±4 | 15.6% | 2026-10-02 | — |
| LMArena |
vision/diagram
|
1,255 ±5.82 | 1,424 ±5 | 14.8% | 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
B Writing 1,389 [1,378–1,400] Percentile 60 · #107 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,367 ±8.61 | 1,394 ±9 | 58.8% | 2026-10-02 | — |
| LMArena |
text/longer_query
|
1,420 ±6.56 | 1,419 ±6 | 33.8% | 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
B Language: 繁體中文 1,405 [1,388–1,423] Percentile 61 · #88 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,462 ±12 | 1,423 ±10 | 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
B Language: English 1,404 [1,395–1,413] Percentile 65 · #97 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,429 ±4.73 | 1,418 ±5 | 93.5% | 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
C Language: Español 1,402 [1,378–1,425] Percentile 47 · #79 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,427 ±14 | 1,424 ±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: 한국어 1,409 [1,386–1,432] Percentile 61 · #54 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/korean
|
1,395 ±17 | 1,432 ±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
C Language: Polski 1,404 [1,384–1,424] Percentile 46 · #71 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,419 ±14 | 1,424 ±11 | 91.0% | 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 1,398 [1,381–1,415] Percentile 59 · #96 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,409 ±10 | 1,415 ±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
History
The same in a table
| Model | 2026-10 |
|---|---|
| Qwen3 VL 235B A22B Instruct | 1,394 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.