Tier 1 — floor
5/5
11 tokens · 1.778s (representative seed)
Tier 2 — restraint (refusal)
5/5
16 tokens · 0.489s (representative seed)
Tier 3 — distractor resistance
5/5
19 tokens · 0.196s (representative seed)
Tier 4 — multi-passage synthesis
0/5
43 tokens · 0.301s (representative seed)
Tier 5 — precise extraction
5/5
37 tokens · 0.303s (representative seed)
Tier 6 — long-context needle
5/5
54 tokens · 0.438s (representative seed)