FrontierSWE: PostgreSQL 18 on SQLite (Mean@5)
Code · 2026-04-16
Build a PostgreSQL 18 server in Zig that uses SQLite for storage. FrontierSWE Mean@5 is the average percentage of verifier tests passed across up to five independent long-horizon agent runs.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Fable 5 | 17.0 |
| GLM-5.2 | 15.0 |
| Opus 4.8 | 13.0 |
| Kimi K2.6 | 13.0 |
| Composer 2.5 | 12.0 |
| Opus 4.7 | 11.0 |
| Grok 4.5 | 11.0 |
| GPT-5.5 | 9.9 |
| DeepSeek-V4-Pro | 8.7 |
| GPT-5.4 | 8.7 |
| GLM-5.1 | 5.7 |
| Gemini 3.1 Pro Preview | 4.2 |
| Opus 4.6 | 3.2 |
| Kimi K2.5 | 0.0 |
| Qwen3.6 Plus (2026-04-02) | 0.0 |