List your company

Open-domain long-horizon evaluation on BrowseComp-Plus (BCP) (test)

10.6AccuracyHiMPO
1.7084.01656.3258.6335Jun 15, 2026

Evaluation results

MethodLinks
HiMPOContext budget=8k ctx., Compression strategy=mem_aware2026.0610.6
HiMPOContext budget=8k ctx., Compression strategy=naive_recency2026.0610.36
HiMPOContext budget=16k ctx., Compression strategy=mem_aware2026.069.76
HiMPOContext budget=4k ctx., Compression strategy=mem_aware2026.069.4
HiMPOContext budget=16k ctx., Compression strategy=naive_recency2026.069.28
HiMPOContext budget=full ctx.2026.069.16
SUPOContext budget=8k ctx., Compression strategy=mem_aware2026.068.19
SUPOContext budget=8k ctx., Compression strategy=naive_recency2026.068.07
MEMPOContext budget=8k ctx., Compression strategy=mem_aware2026.067.95
MEMPOContext budget=8k ctx., Compression strategy=naive_recency2026.067.71
SFT onlyContext budget=8k ctx., Compression strategy=mem_aware2026.067.59
ReActContext budget=4k ctx., Compression strategy=mem_aware2026.067.47
SFT onlyContext budget=8k ctx., Compression strategy=naive_recency2026.067.47
SUPOContext budget=16k ctx., Compression strategy=mem_aware2026.067.35
SUPOContext budget=4k ctx., Compression strategy=mem_aware2026.067.23
SUPOContext budget=16k ctx., Compression strategy=naive_recency2026.067.11
MEMPOContext budget=4k ctx., Compression strategy=mem_aware2026.066.99
MEMPOContext budget=16k ctx., Compression strategy=mem_aware2026.066.99
SUPOContext budget=full ctx.2026.066.99
SFT onlyContext budget=4k ctx., Compression strategy=mem_aware2026.066.75
SFT onlyContext budget=16k ctx., Compression strategy=mem_aware2026.066.75
MEMPOContext budget=16k ctx., Compression strategy=naive_recency2026.066.75
MEMPOContext budget=full ctx.2026.066.63
SFT onlyContext budget=16k ctx., Compression strategy=naive_recency2026.066.51
ReActContext budget=16k ctx., Compression strategy=naive_recency2026.066.39
ReActContext budget=8k ctx., Compression strategy=mem_aware2026.066.27
SFT onlyContext budget=full ctx.2026.066.27
ReActContext budget=8k ctx., Compression strategy=naive_recency2026.065.3
ReActContext budget=16k ctx., Compression strategy=mem_aware2026.065.3
ReActContext budget=4k ctx., Compression strategy=naive_recency2026.065.18
ReActContext budget=full ctx.2026.065.06
HiMPOContext budget=4k ctx., Compression strategy=naive_recency2026.064.7
SUPOContext budget=4k ctx., Compression strategy=naive_recency2026.062.53
MEMPOContext budget=4k ctx., Compression strategy=naive_recency2026.062.29
SFT onlyContext budget=4k ctx., Compression strategy=naive_recency2026.062.05