EdgeBench

What is EdgeBench?

EdgeBench evaluates autonomous agents on 134 real-world tasks across scientific discovery, software engineering, optimization, knowledge work, formal…

Released

2026-07-02

Evaluates

Agents, Agents & Tool Use, Environment Learning, Environment learning, General AI, Learning from feedback, Long-horizon task execution, Mathematics & Formal Science, Scaling Laws, Scientific Research & AI for Science, Software & AI Compute

Openness

unknown

Importer

Claire Radar

Review status

source-reviewed

Reported scores

0

Source provenance

• Original evidence https://arxiv.org/abs/2607.05155

EdgeBench paper, code and dataset

• Paper https://arxiv.org/abs/2607.05155
• Code https://github.com/ByteDance-Seed/EdgeBench
• Dataset https://huggingface.co/datasets/ByteDance-Seed/EdgeBench
• Website https://edge-bench.org/

Which sources cite EdgeBench?

• EdgeBench

Related Agents, Agents & Tool Use, Environment Learning, Environment learning, General AI, Learning from feedback, Long-horizon task execution, Mathematics & Formal Science, Scaling Laws, Scientific Research & AI for Science, Software & AI Compute benchmarks

• $\tau^\tau$-Bench: An Environment for End-To-End, Realistic Agent Construction
• 3DCodeBench
• 5-Dialects-BN
• A3S-Bench
• AARRI-Bench
• ActBench
• Active-SWE
• ActiveSciBench