Codesota · Models · Agent Q (GPT-4o)MultiOn1 results · 1 benchmarks
Model card

Agent Q (GPT-4o).

MultiOnproprietaryUnknown paramsMCTS + DPO self-play web agent on GPT-4o

Agent Q uses Monte Carlo Tree Search + direct preference optimization via self-play. arxiv:2408.07199.

§ 01 · Benchmarks

Every benchmark Agent Q (GPT-4o) has a recorded score for.

#BenchmarkArea · TaskMetricValueRankDateSource
01WebArenaAgentic AI · Web & Desktop Agentssuccess-rate50.5%#4/62023-07-26source ↗
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 02 · Strengths by area

Where Agent Q (GPT-4o) actually performs.

Agentic AI
1
benchmark
avg rank #4.0
§ 03 · Papers

1 paper with results for Agent Q (GPT-4o).

  1. 2023-07-26· Agentic AI· 1 result

    WebArena: A Realistic Web Environment for Building Autonomous Agents

§ 05 · Sources & freshness

Where these numbers come from.

arxiv
1
result
1 of 1 rows marked verified.