Nebius Releases LongTail-SWE Benchmark for 13 Underrepresented Languages

Nebius released LongTail-SWE, a benchmark of 202 real repository-level tasks across 13 underrepresented programming languages, where Qwen3.8 solved 43% of tasks.

2026-11-10 ~ 2026-11-11 · 3 related posts

• Nebius releases Longtail-SWE: 202 repo-level SWE tasks across 13 underrepresented languages — nebius · 2026-11-10
• LongTail-SWE: 202 real tasks across 13 underrepresented languages, Qwen3.8 solves 43% vs Qwen3.6's 17% — Fabulous_Pollution10 · 2026-11-10
• Longtail-SWE benchmark ships 202 real-world SWE tasks across 13 underrepresented languages — _akhaliq · 2026-11-11