GLM-4.5-Air is the lightweight variant of our latest flagship model family, also purpose-built for agent-centric applications. Like GLM-4.5, it adopts the Mixture-of-Experts (MoE) architecture but with a more compact parameter...
text->text| Provider | Region / Tag | Context | Input/M | Output/M | Uptime 1d | Latency | Quant | Cache |
|---|---|---|---|---|---|---|---|---|
| SiliconFlow | siliconflow/fp8 | 131K | $0.140 | $0.860 | 99.41% | - | fp8 | - |
| Novita | novita/bf16 | 131K | $0.130 | $0.850 | 99.32% | - | bf16 | - |
| Z.AI | z-ai/fp8 | 131K | $0.200 | $1.10 | 99.10% | - | fp8 | - |
| Rank | Category | Variant | Trend | Elo | Votes |
|---|---|---|---|---|---|
| #108 | Japanese | - | 1327 | 785 | |
| #120 | French | - | 1406 | 468 | |
| #129 | Chinese | - | 1414 | 1,615 | |
| #133 | Industry Mathematical | - | 1391 | 1,453 | |
| #133 | Math | - | 1390 | 1,540 | |
| #143 | Industry Life And Physical And Social Science | - | 1397 | 4,931 | |
| #143 | Spanish | - | 1363 | 723 | |
| #144 | Expert | - | 1390 | 1,389 | |
| #144 | Industry Software And It Services | - | 1415 | 10,657 | |
| #145 | Coding | - | 1427 | 6,104 | |
| #145 | German | - | 1346 | 672 | |
| #145 | Industry Business And Management And Financial Operations | - | 1372 | 5,506 | |
| #146 | English | - | 1390 | 14,824 | |
| #146 | Longer Query | - | 1379 | 6,635 | |
| #147 | Industry Medicine And Healthcare | - | 1396 | 1,645 | |
| #148 | Hard Prompts English | - | 1403 | 7,654 | |
| #150 | Hard Prompts | - | 1391 | 14,876 | |
| #150 | Overall | - | 1373 | 31,077 | |
| #151 | Instruction Following | - | 1363 | 8,116 | |
| #151 | Multi Turn | - | 1369 | 5,122 | |
| #153 | Exclude Ties | - | 1343 | 21,869 | |
| #153 | Korean | - | 1284 | 637 | |
| #153 | Russian | - | 1357 | 1,485 | |
| #160 | Industry Entertainment And Sports And Media | - | 1326 | 5,697 | |
| #161 | Industry Writing And Literature And Language | - | 1342 | 7,024 | |
| #162 | Creative Writing | - | 1328 | 4,021 | |
| #162 | Industry Legal And Government | - | 1374 | 1,932 |