BREAKING
AA-Briefcase Benchmark Launches
Multi-week knowledge-work tasks
14 scenarios
291 tasks
3Real deliverables
AA-Briefcase Elo leaderboard
Claude Fable 51587
Opus 4.81356
GLM-5.21266
GPT-5.51159
0B
total params
0B
active params
0M
context tokens
0x
vs GLM-5.1
0x
vs Kimi K2.6
0%
tasks fully solved
Real knowledge-work stays hard
AI NEWS BLITZ
Artificial Analysis just launched AA-Briefcase, a new test for real knowledge-work.