Unverified50% confidenceBenchmarkExact time
In the WebShop decision-making task, ReAct scored 66.6, surpassing the imitation learning baseline but falling short of the human expert score of 82.1.
1
Sources
50%
Confidence
Long-term
Relevance
9/9/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedBenchmarks such as WebArena and Mind2Web show that the current strongest web agents still achieve a success rate of less than 50% on complex web tasks.68% similarUnverifiedExperiments by the Long-Term team showed that using the same model with only adjustments to rule constraints and the memory system raised the score from 52.8% to 66.5%, jumping the ranking from Top 30 to Top 5.65% similarUnverifiedMultiple benchmark tests in 2023 (such as the HaHackathon dataset) showed that top LLMs still have significantly lower accuracy than ordinary human annotators on sarcasm recognition tasks63% similarUnverifiedOn the BrowseWebApp Bench benchmark, the accuracy of the fine-tuned model improved from 22% to 63%.62% similarUnverifiedIn the same reproduction experiment, ResNet-20 achieved a test accuracy of 90.4% and ResNet-56 achieved a test accuracy of 91.7%, with training accuracies of 97.4% and 99.0% respectively.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/884336API
curl https://kongchang.com/api/v1/knowledge/claims/884336MCP
get_claim(id=884336)