Hieroglyphic Benchmark

Tools & Infrastructure

Hieroglyphic Benchmark is a lateral and abstract-reasoning test for AI models. The leaked Gemini 3.5 "Snowbunny" achieved an 80% score on it.

🌐 Visit website ↗

🔗 Similar tools

Stripe Agent Benchmark

Stripe Agent Benchmark is an end-to-end evaluation suite that tests whether AI agents can …
GPQA Diamond ★ 532

GPQA Diamond is a benchmark dataset of expert-level, graduate-school science questions use…
GLM-4 ★ 7.1k

GLM-4 is an AI model achieving a 77.8% score on the SWE-bench benchmark, with the lowest h…

Models & AI Labs
BrowserGym ★ 1.3k

BrowserGym provides a Gym-style benchmark and environment toolkit for evaluating web agent…
LLM-Agent-Benchmark-List ★ 170

A curated list of benchmarks for evaluating large language models, covering tasks like rea…
Agents' Last Exam (ALE)

ALE is a UC Berkeley (RDI) and Snorkel AI benchmark from June 2026 that evaluates agents o…