The Qwen3.5 122B-A10B native vision-language model is built on a hybrid architecture that integrates a linear attention mechanism with a sparse mixture-of-experts model, achieving higher inference efficiency. In terms of...
text+image+video->text| Provider | Region / Tag | Context | Input/M | Output/M | Uptime 1d | Latency | Quant | Cache |
|---|---|---|---|---|---|---|---|---|
| Alibaba | alibaba | 262K | $0.260 | $2.08 | 99.30% | - | - | - |
| Novita | novita/bf16 | 262K | $0.400 | $3.20 | 98.95% | - | bf16 | - |
| AtlasCloud | atlas-cloud/fp8 | 262K | $0.300 | $2.40 | 98.85% | - | fp8 | - |
| DeepInfra | deepinfra/fp4 | 262K | $0.290 | $2.40 | 94.48% | - | fp4 | - |
| SiliconFlow | siliconflow/fp8 | 262K | $0.260 | $2.08 | 72.01% | - | fp8 | - |
| Rank | Category | Variant | Trend | Elo | Votes |
|---|---|---|---|---|---|
| #73 | French | - | 1451 | 868 | |
| #86 | German | - | 1414 | 422 | |
| #88 | Industry Business And Management And Financial Operations | - | 1422 | 5,448 | |
| #89 | Math | - | 1424 | 1,778 | |
| #91 | Chinese | - | 1462 | 1,635 | |
| #91 | Spanish | - | 1411 | 855 | |
| #92 | Industry Mathematical | - | 1430 | 1,581 | |
| #92 | Japanese | - | 1363 | 229 | |
| #93 | Expert | - | 1445 | 2,476 | |
| #99 | Industry Life And Physical And Social Science | - | 1437 | 4,629 | |
| #100 | Non English | - | 1401 | 15,071 | |
| #101 | English | - | 1431 | 13,434 | |
| #102 | Instruction Following | - | 1407 | 9,024 | |
| #102 | Korean | - | 1349 | 451 | |
| #103 | Hard Prompts English | - | 1444 | 8,837 | |
| #104 | Exclude Ties | - | 1405 | 20,864 | |
| #104 | Overall | - | 1417 | 28,523 | |
| #105 | Industry Legal And Government | - | 1425 | 2,215 | |
| #105 | Multi Turn | - | 1418 | 4,951 | |
| #107 | Coding | - | 1459 | 7,807 | |
| #108 | Industry Software And It Services | - | 1451 | 11,194 | |
| #109 | Hard Prompts | - | 1433 | 17,856 | |
| #113 | Industry Writing And Literature And Language | - | 1389 | 6,459 | |
| #113 | Polish | - | 1384 | 585 | |
| #114 | Longer Query | - | 1418 | 10,930 | |
| #118 | Industry Medicine And Healthcare | - | 1433 | 2,067 | |
| #122 | Russian | - | 1397 | 3,011 | |
| #129 | Creative Writing | - | 1367 | 4,349 | |
| #129 | Industry Entertainment And Sports And Media | - | 1367 | 5,491 |
| Rank | Category | Variant | Trend | Elo | Votes |
|---|---|---|---|---|---|
| #36 | Chinese | - | 1280 | 744 | |
| #37 | Humor | - | 1226 | 417 | |
| #45 | Creative Writing Vision | - | 1228 | 741 | |
| #45 | English | - | 1227 | 5,313 | |
| #45 | Ocr | - | 1240 | 8,747 | |
| #45 | Overall | - | 1228 | 12,519 | |
| #47 | Homework | - | 1252 | 1,882 | |
| #49 | Diagram | - | 1245 | 3,245 | |
| #50 | Creative Writing | - | 1208 | 741 |
| Rank | Category | Variant | Trend | Elo | Votes |
|---|---|---|---|---|---|
| #60 | Webdev-react | - | 1359 | 7,145 | |
| #63 | Webdev | - | 1364 | 8,240 | |
| #64 | Overall | - | 1364 | 8,240 | |
| #72 | Webdev-html | - | 1347 | 994 |