Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Multimodal tool-use on GTA

67.66Answer Accuracy

ReGRPO (default, λval = 0)

11.957626.418840.8855.3412Apr 30, 2025Jul 10, 2025Sep 19, 2025Nov 29, 2025Feb 8, 2026Apr 20, 2026Jun 30, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
67.6676.3593.77
2026.03
60.968.398.7
2025.04
60.2672.4191.87
2026.06
60.2672.4191.87
2026.03
58.465.194.3
2025.04
57.6956.1100
2026.06
57.6956.1100
2026.03
57.163.495.1
2025.04
57.0563.4195.12
2026.06
57.0563.4195.12
2026.03
56.765.787.4
2025.04
53.8564.6384.32
2026.06
53.8564.6384.32
2026.03
53.864.684.3
2025.04
52.5665.8580.49
2026.06
52.5665.8580.49
2026.03
49.256.874.2
2026.03
46.855.469.8
2025.04
46.59--
2026.06
46.59--
2026.03
44.250.669.1
2025.04
42.3144.8565.19
2026.06
42.3144.8565.19
2025.04
41.52--
2026.06
41.52--
2025.04
33.9736.5956.1
2026.06
33.9736.5956.1
2025.04
32.0536.7552.18
2026.06
32.0536.7552.18
2026.03
23.249.457.2
2026.03
14.114.925.1
2025.04
14.114.9725.08
2026.06
14.114.9725.08