JourneyBench (Dynamic-Prompt Agent) - Failing Function: leaderboard
The Failing Function scenarios of JourneyBench (Balaji et al., 2026), a customer-support benchmark of business workflow graphs: one tool call fails and the agent must stop the dependent steps. In this setting the authors' orchestrator walks the procedure graph one node at a time and swaps the prompt and tools at each step. User Journey Coverage Score from 0 to 1 (a conversation scores 0 on any missing, extra or misordered tool call and otherwise its share of correct tool parameters); higher is better.
Metric: User Journey Coverage Score (0-1; a conversation scores 0 on any missing, extra or misordered tool call, otherwise its tool-parameter accuracy; GPT-4o simulated user, 40-turn limit; Failing Function scenarios; the authors' orchestrator walks the SOP graph node by node and swaps the prompt and tools at each transition). Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 0.86 | #741 |
| 2 | GPT-4o Mini | 0.82 | #1627 |
| 3 | Claude 3.5 Haiku | 0.78 | #1563 |
| 4 | Llama 3.3 70B Instruct | 0.35 | #1353 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=journeybench-dynamic-prompt-agent-failing-function · How It Works · Data refreshed daily, snapshot 2026-11-12.