QCalEval - Technical Description: leaderboard
Metric: Score (0-100) for a structured JSON description of plot type, axes and salient features (half exact-match fields, half key-point coverage judged by GPT-5.4 and Gemini 3.1 Pro Preview), on the 243 quantum-calibration plots of QCalEval (22 experiment families, superconducting qubits and neutral atoms), zero-shot with the family background text, greedy decoding; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 90.9 |
| 2 | Claude Opus 4.6 | 90.8 |
| 3 | GPT-5.4 Mini | 90.3 |
| 4 | Claude Sonnet 4.6 | 89.7 |
| 5 | Gemini 3.1 Flash Lite | 89.2 |
| 6 | Gemini 3.1 Pro (Preview) | 88.5 |
| 7 | Qwen 3.5 397B A17B | 88.1 |
| 8 | Qwen 3.5 27B | 87 |
| 9 | Qwen 3.5 35B A3B | 86.8 |
| 10 | Qwen 3.5 122B A10B | 86.6 |
| 11 | Gemma 4 31B (IT) | 85.6 |
| 12 | Claude Haiku 4.5 | 83.4 |
| 13 | Qwen 3.5 9B | 81.5 |
| 14 | InternVL3-38B | 79.2 |
| 15 | InternVL3-78B | 76.3 |
Interactive version: theaggregate.ai/benchmark?slug=qcaleval-technical-description · How It Works · Data refreshed daily, snapshot 2026-10-07.