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

Interactive Agent Task on TRIP-Bench (Hard LIT)

4,400Loose Success

GPT-5.2

Updated 3mo ago

Evaluation Results

Method	Links
GPT-5.2 2026.02		4,400	1,400
DeepSeek-V3.2 2026.02		3,600	200
Claude-Sonnet-4.5 2026.02		2,800	0
DeepSeek-V3.2 2026.02		1,600	200
Gemini-3-Pro 2026.02		1,600	0
Gemini-3-Pro 2026.02		1,200	0
GPT-5.2 2026.02		1,000	0
Claude-Sonnet-4.5 2026.02		1,000	0
GLM-4.7 2026.02		1,000	0
Gemini-3-Flash 2026.02		1,000	0
Gemini-3-Flash 2026.02		800	0
GLM-4.7 2026.02		600	0
Gemini-3-Flash 2026.02		26	0
DeepSeek-V3.2 2026.02		26	0
Claude-Sonnet-4.5 2026.02		22	0
GPT-5.2 2026.02		20	0
Gemini-3-Pro 2026.02		18	0
Gemini-3-Flash 2026.02		16	0
Gemini-3-Pro 2026.02		12	0
Claude-Sonnet-4.5 2026.02		10	0
GPT-5.2 2026.02		8	0
Qwen3-235B-A22B-Instruct-2507 2026.02		2	0
DeepSeek-V3.2 2026.02		2	0
Kimi-K2-0905-Preview 2026.02		0	0
Qwen3-235B-A22B-Instruct-2507 2026.02		0	0
Qwen3-235B-A22B-Thinking-2507 2026.02		0	0
Kimi-K2-Thinking 2026.02		0	0
Kimi-K2-0905-Preview 2026.02		0	0
GLM-4.7 2026.02		0	0
Qwen3-235B-A22B-Thinking-2507 2026.02		0	0
Kimi-K2-Thinking 2026.02		0	0
GLM-4.7 2026.02		0	0