Neura Intelligence Index

AI Model Benchmarks

Back to Leaderboard

OSWorld

Real-world computer task completion in operating system environments.

agent0-100 percent

Model Rankings (20)

#	Model	Score	Provider	Source
	B Seed 2.1 Pro🇨🇳			verified
	B Seed 2.1 Turbo🇨🇳			verified
	Claude Opus 4.6🇺🇸			verified
	Claude Sonnet 4.6🇺🇸			verified
	Q Qwen3 VL 235B A22B Instruct🇨🇳			verified
	Claude Opus 4.5🇺🇸			verified
	Z GLM-5V-Turbo🇨🇳			verified
	Claude Sonnet 4.5🇺🇸			verified
	Claude Haiku 4.5🇺🇸			verified
	Q Qwen3 VL 32B Thinking🇨🇳			verified
	Q Qwen3 VL 235B A22B Thinking🇨🇳			verified
	Q Qwen3 VL 8B Instruct🇨🇳			verified
	Q Qwen3 VL 8B Thinking🇨🇳			verified
	Q Qwen3 VL 32B Instruct🇨🇳			verified
	Q Qwen3 VL 4B Thinking🇨🇳			verified
	Q Qwen3 VL 30B A3B Thinking🇨🇳			verified
	Q Qwen3 VL 30B A3B Instruct🇨🇳			verified
	Q Qwen3 VL 4B Instruct🇨🇳			verified
	Q Qwen2.5 VL 72B Instruct🇨🇳			verified
	Q Qwen2.5 VL 32B Instruct🇨🇳			verified

OSWorld — AI Model Leaderboard & Results | Neura Market