QCalEval: leaderboard
Metric: Mean score (0-100) over the six question types (description, outcome, significance, fit, parameters, diagnosis), 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 | Gemini 3.1 Pro (Preview) | 72.3 |
| 2 | Gemma 4 31B (IT) | 68.8 |
| 3 | Gemini 3.1 Flash Lite | 68.2 |
| 4 | Claude Opus 4.6 | 67.8 |
| 5 | Claude Sonnet 4.6 | 66.5 |
| 6 | GPT-5.4 | 64.6 |
| 7 | Qwen 3.5 397B A17B | 58.6 |
| 8 | Qwen 3.5 27B | 58.5 |
| 9 | Qwen 3.5 122B A10B | 57.1 |
| 10 | GPT-5.4 Mini | 55.7 |
| 11 | Qwen 3.5 35B A3B | 55.5 |
| 12 | Qwen 3.5 9B | 53 |
| 13 | Claude Haiku 4.5 | 50.5 |
| 14 | InternVL3-78B | 48.2 |
| 15 | InternVL3-38B | 44.1 |
Interactive version: theaggregate.ai/benchmark?slug=qcaleval · How It Works · Data refreshed daily, snapshot 2026-10-07.