QCalEval - Significance Analysis: leaderboard
Metric: Score (0-100) for the scientific significance analysis, three expert key points per item scored 0, 0.5 or 1 by GPT-5.4 and Gemini 3.1 Pro Preview and averaged, 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: Around January 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 65.5 |
| 2 | GPT-5.4 | 63.7 |
| 3 | Claude Sonnet 4.6 | 63.4 |
| 4 | Gemini 3.1 Pro (Preview) | 61.1 |
| 5 | Gemma 4 31B (IT) | 59.8 |
| 6 | Gemini 3.1 Flash Lite | 59.4 |
| 7 | Qwen 3.5 397B A17B | 52 |
| 8 | Qwen 3.5 122B A10B | 49 |
| 9 | Qwen 3.5 27B | 48.3 |
| 10 | GPT-5.4 Mini | 48.3 |
| 11 | Qwen 3.5 35B A3B | 45.7 |
| 12 | Claude Haiku 4.5 | 40.8 |
| 13 | Qwen 3.5 9B | 39.5 |
| 14 | InternVL3-78B | 34.1 |
| 15 | InternVL3-38B | 27.6 |
Interactive version: theaggregate.ai/benchmark?slug=qcaleval-significance-analysis · How It Works · Data refreshed daily, snapshot 2026-10-07.