zhipu-ai

Checked numbers

Each one tied to the claim it was checked against, in the words it was read from.

• glm-5-3 Benchmark score: Binary Exploitation (Anthropic internal, 100 tasks) · measured by others · 1 Oct 4%