IC-221GPT-4o ReAct success rate drops from 47% on synchronous to 11% on asynchronous planning tasks, and all other tested LLMs show equal or worse performance

Gonzalo Gonzalez-Pumariega, Leong Su Yean, Neha Sunkara, Sanjiban Choudhury

SourceRobotouille: An Asynchronous Planning Benchmark for LLM Agents

The paper evaluates multiple LLMs using three agent baselines (I/O, I/O CoT, ReAct) on the Robotouille benchmark's synchronous and asynchronous cooking-planning datasets. GPT-4o with ReAct is the best performer, achieving 47% on synchronous tasks but only 11% on asynchronous tasks. GPT-4o-mini drops from 11% to 0%, Gemini-1.5-Flash scores 0% on both, and Claude-3-Haiku scores 2% on both. Open-source models (Qwen2-72B, Llama-3.1-70B, Gemma-2-27B) score at most 7% on synchronous and 2% on asynchronous. The gap is confirmed by a controlled comparison where the first three async tasks are made synchronous (delay set to 0), and GPT-4o ReAct improves from 20-40% to 50-60%.

Evidence
correlational
Key metric
GPT-4o ReAct: 47.0% sync, 11.0% async; GPT-4o-mini ReAct: 11.0% sync, 0.00% async; Gemini-1.5-Flash ReAct: 0.00% both; Claude-3-Haiku ReAct: 2.00% both; Qwen2-72B-Instruct: 7.00% sync, 2.00% async; Meta-Llama-3.1-70B-Instruct: 2.00% sync, 0.00% async
Caveat
Success is defined as reaching the goal within 1.5 times the optimal number of steps; the benchmark has only 10 tasks per dataset with 10 procedurally generated instances each, and a single in-context example is provided.
Model
GPT-4o mini, Gemini 1.5 / Gemini Pro 1.5 Gemini 1.5 Flash, Claude 3 Haiku, Qwen 2 Qwen2-72B-Instruct, Qwen2-32B-Instruct, Llama 3.1 Meta-Llama-3.1-70B-Instruct, Gemma 2 Gemma-2-27B-IT, Gemma-2-9B-IT
Concepts
Failure mode
Methods
ReAct [primary], I/O CoT [compared-to]
Related work
ReAct [builds-on], Chain-of-Thought prompting / Chain-of-Thought (CoT) prompting / CoT prompting / Few-shot Chain of Thought / Few-shot CoT / Chain-of-Thought (CoT-S) / CoT-bag prompting / Few-shot CoT prompting / Rationale prompting / Zero-shot CoT prompting / Wei et al. (2022) chain-of-thought prompting [builds-on]
Related findings
IC-222, IC-223, IC-224
Extraction
automatic-extraction