scipy: outdated calls AI models write
2 corrections for scipy 1.18.1 (PyPI): code that gemini-3.8-flash and gpt-6.1-sol wrote in our sandbox runs for 1.17.1 and earlier, which fails on 1.18.1. Each one is verified on both versions.
- Corrections
- 2
- Fails on
- 1.18.1
- Works on
- 1.17.1
- Released
- 2026-08-21
| Changed API | Change | Models wrote | Correct | Error | Seen in | Worked |
|---|---|---|---|---|---|---|
| scipy.optimize.nnls | removed | x, residual_norm = nnls(A, b, atol=1e-8) | x, residual_norm = nnls(A, b) | TypeError | gpt-6.1-sol | 2/2 |
| scipy.linalg.sqrtm | changed | root, errest = linalg.sqrtm(A, disp=False) | root = linalg.sqrtm(A) | TypeError | gemini-3.8-flash, gpt-6.1-sol | 4/4 |
Wide table: scroll it sideways to see every column.
| Model | Answers run | Outdated code found | Share |
|---|---|---|---|
| claude-sonnet-5-5 | 12 | 2 | 17 % |
| gpt-6.1-sol | 12 | 6 | 50 % |
| gemini-3.8-flash | 12 | 4 | 33 % |
For agents
- This page as JSON:
https://formachines.ai/errata/pypi/scipy/index.json - Whole database in plain text:
/errata/llms-full.txt - Install the skill:
npx skills add formachines-ai/skills@model-errata(source on GitHub) - Confirm, dispute or add a correction: feedback form and API
- Apply a correction on your own only when its
proof.auto_applyis true (Automatic application is allowed only when proof.equivalent is true and proof.applicability.status is applies; otherwise show the correction to the user as a lead and do not apply it on your own.)