vCX-Hard - Tool Calling (Reasoning): leaderboard
vCX-Hard tool-call accuracy in each model's best configuration with reasoning enabled. Higher is better. Source: retellai.com/blog/vcx-hard-benchmark.
Metric: Tool-Call Accuracy (%). Source: www.retellai.com. Saturation forecast: Around February 2027. 27 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 88.3 | #33 |
| 2 | Gemini 3.1 Pro (Preview) | 85.9 | #65 |
| 3 | Claude Sonnet 4.5 | 85.8 | #148 |
| 4 | Gemini 3.5 Flash | 84 | #67 |
| 5 | GLM-5.1 | 83.8 | #109 |
| 6 | GPT-5.1 | 83.7 | #145 |
| 7 | GPT-5 | 83.2 | #106 |
| 8 | Gemini 3 Flash | 81.8 | #105 |
| 9 | GPT-5.2 | 81.5 | #114 |
| 10 | Gemini 2.5 Pro | 80.6 | #159 |
| 11 | Claude Sonnet 4.6 | 80.4 | #96 |
| 12 | Claude Haiku 4.5 | 79.4 | #288 |
| 13 | GPT-5.4 | 77.4 | #84 |
| 14 | Gemini 3.1 Flash Lite | 75.8 | #204 |
| 15 | GPT-5 Mini | 75.6 | #192 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=vcx-hard-tool-calling-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-06.