ARFBench: leaderboard
Datadog's Agentic Reasoning & Function-calling benchmark, scoring LLMs on multi-step tool-use trajectories in realistic observability and operations workflows.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around May 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 | 62.7 | #98 |
| 2 | GPT-5.4 | 61.3 | #79 |
| 3 | Gemini 3 Pro | 58.1 | #80 |
| 4 | Claude Opus 4.6 | 54.8 | #63 |
| 5 | GPT-4.1 | 47.9 | #248 |
| 6 | GPT-4o | 47.2 | #343 |
| 7 | Claude Sonnet 4.5 | 47.2 | #142 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=arfbench · How It Works · Data refreshed daily, snapshot 2026-10-20.