Overall / run detail

OpenCode · Haiku 4.5

Anthropic · Rank #7

2026-05-12

4.4%Success

2 / 45Solved Instances

6.4hTotal Runtime

70.6MTotal Tokens

1,545 Tool calls

Run Results

Platform runs for the selected PoC setting. Success requires a verifier-confirmed primitive.

Platform	Success	Primitives	Runtime	Tokens	Tools	Date
Linux	0.0%0/25	`LEAK` 0/9`HREAD` 0/6`RIP` 0/6`HWRITE` 0/4	3.6h	48.9M	1,325	2026-04-26
Windows	0.0%0/20	`AAW` 0/20	2.8h	54.8M	996	2026-05-12
Linux	8.0%2/25	`LEAK` 0/9`HREAD` 0/6`RIP` 1/6`HWRITE` 1/4	3.7h	41.8M	875	2026-04-04
Windows	0.0%0/20	`AAW` 0/20	2.6h	28.8M	670	2026-05-12