| multimodal-reasoning-on-babyvision | GLM-5.3-Flash | #1 | Accuracy: 53.4 |
| agents-on-toolathlon-verified-108-tasks | GLM-5.3-Flash | #2 | Pass@1: 0.784 |
| computer-use-agents-on-osworld-2-0 | GLM-5.3-Flash | #2 | Success Rate: 59.1 |
| video-question-answering-on-mvbench | GLM-5.3-Flash | #2 | Avg.: 77.8 |
| video-question-answering-on-mmvu | GLM-5.3-Flash | #3 | Accuracy: 80.5 |
| agents-on-agents-last-exam-ale-cli | GLM-5.3-Flash (Claude Code) | #5 | Pass Rate: 26.3 |
| agents-on-automationbench-public | GLM-5.3-Flash | #5 | Success Rate: 48.8 |
| code-generation-on-nl2repo | GLM-5.3-Flash | #6 | Accuracy: 56.3 |
| code-generation-on-terminal-bench-2-1 | GLM-5.3-Flash (Claude Code) | #11 | Accuracy: 84.3 |
| code-generation-on-deepswe | GLM-5.3-Flash (mini-swe-agent) | #16 | Pass@1: 0.634 |
| sentence-completion-on-hellaswag | GLM-5.3-Flash-Base | #27 | Accuracy: 87.1 |