{"asset_base":"https://tongyi-mai.github.io/OpenGCUA-blog/","benchmark":"OSWorld-v2 v2.1","leaderboard_tasks":82,"paper":"https://arxiv.org/abs/2610.22200","models":[{"id":"gpt-6-astra_all","name":"GPT-6 Astra","task_count":108,"record_count":179,"index":"trajectory-data/gpt-6-astra_all.json","zip":"gpt-6-astra_all/traj.json.zip","zip_bytes":4315200,"zip_sha256":"6ad3b6e964df2bf7f748adb7170954d195f940de9a3e02e0dd9696d5e0242700","score":75.7,"baseline_score":67.1,"success":46.3,"steps":33.5,"tokens":31.9},{"id":"opus-5_all","name":"Claude Opus 5","task_count":108,"record_count":228,"index":"trajectory-data/opus-5_all.json","zip":"opus-5_all/traj.json.zip","zip_bytes":4809998,"zip_sha256":"0f967c5e69971eca447508c452b94f4e128ae70e7cf447eadafa925420898be7","score":70.1,"baseline_score":60.7,"success":36.6,"steps":61.4,"tokens":89.3},{"id":"gemini-3.8-flash_all","name":"Gemini 3.8 Flash","task_count":108,"record_count":1984,"index":"trajectory-data/gemini-3.8-flash_all.json","zip":"gemini-3.8-flash_all/traj.json.zip","zip_bytes":26226295,"zip_sha256":"e28596724fc0721a1f0e1d8438ef842b1da424aa1e356988ae3cc995aa6c1f5d","score":67.0,"baseline_score":50.1,"success":31.7,"steps":101.2,"tokens":82.0},{"id":"glm-5.3-flash-all","name":"GLM-5.3-Flash","task_count":108,"record_count":137,"index":"trajectory-data/glm-5.3-flash-all.json","zip":"glm-5.3-flash-all/traj.json.zip","zip_bytes":22076219,"zip_sha256":"089d0096b15fe8e435036a4a042aeb245f02e962060839140a25e512397fc595","score":64.9,"baseline_score":29.6,"success":29.3,"steps":75.5,"tokens":131.0},{"id":"qwen3.8-max_all","name":"Qwen3.8-Max","task_count":108,"record_count":123,"index":"trajectory-data/qwen3.8-max_all.json","zip":"qwen3.8-max_all/traj.json.zip","zip_bytes":13877308,"zip_sha256":"f53325b6aee30afd6dea168135a091540c743c64e99f04d5bbc10a038cac96f8","score":64.0,"baseline_score":46.9,"success":28.9,"steps":57.0,"tokens":99.5},{"id":"qwen3.8-flash_all","name":"Qwen3.8-Flash","task_count":108,"record_count":155,"index":"trajectory-data/qwen3.8-flash_all.json","zip":"qwen3.8-flash_all/traj.json.zip","zip_bytes":27454139,"zip_sha256":"2341bcc064a0821733855f143a200f33bab7a7d1c32c7b4ff39d32f058f084e9","score":60.5,"baseline_score":38.9,"success":27.6,"steps":86.2,"tokens":159.6},{"id":"kimi-k3_all","name":"Kimi K3","task_count":108,"record_count":436,"index":"trajectory-data/kimi-k3_all.json","zip":"kimi-k3_all/traj.json.zip","zip_bytes":22917671,"zip_sha256":"0a5d4dfe65077c17e247ddec3d43c9d404516e8bc30277e39e05c5352cfd187e","score":59.2,"baseline_score":40.7,"success":22.0,"steps":81.9,"tokens":139.6},{"id":"deepseek-flash-all","name":"DeepSeek-V4.1-Flash","task_count":108,"record_count":238,"index":"trajectory-data/deepseek-flash-all.json","zip":"deepseek-flash-all/traj.json.zip","zip_bytes":16879904,"zip_sha256":"8616fd2f586f507ebaecaee52a495bead8c6ec659fd8a5456e591ddf23b0c29f","score":57.7,"baseline_score":28.0,"success":26.8,"steps":84.9,"tokens":180.4}]}
