Quantum API Drift - Qiskit 2.0: leaderboard
Part of the quantum-api-drift benchmark (Khalifa University, 2026): code for 50 Qiskit HumanEval-derived tasks is requested for Qiskit 2.0 and executed in a Qiskit 2.0 environment, three samples per prompt at temperature 0.8. The score is Pass@1 in percent; higher is better.
Metric: Pass@1 (%; unbiased estimator from 3 samples per task, macro-averaged over 50 Qiskit HumanEval-derived tasks) of code generated for Qiskit 2.0 (2.0.0), executed in that Qiskit environment; REST API access at temperature 0.8 with a 1,024-token output cap; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 16 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 85.33 | #79 |
| 2 | GPT-5.3 Codex | 79.84 | #116 |
| 3 | Claude Sonnet 4.6 | 70 | #161 |
| 4 | GPT-5.4 | 70 | #123 |
| 5 | Grok 4.20 | 63.33 | #297 |
| 6 | Kimi K2 | 50 | #439 |
| 7 | Nemotron 3 Super | 49.33 | #768 |
| 8 | Gemma 4 31B (IT) | 47.33 | #342 |
| 9 | GPT-OSS-120B | 42.67 | #700 |
| 10 | GPT-5.4 Nano | 40 | #635 |
| 11 | Devstral 2 | 33.11 | #1003 |
| 12 | Qwen3 Coder | 31.33 | #586 |
| 13 | Qwen 3 32B | 29.33 | #960 |
| 14 | GLM-4.7 Flash | 24.67 | #1259 |
| 15 | Llama 3.3 70B Instruct | 12.67 | #1357 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=quantum-api-drift-qiskit-2-0 · How It Works · Data refreshed daily, snapshot 2026-11-13.