4B model Spark-X2.6 scores 94.6/100 on community benchmark, beating 27B-35B models
mesolimbic-bliss · reddit · 2026-10-31
A Reddit user benchmarked the Spark-X2.6-4B GGUF model on their community-run ShortBench platform and got 94.6/100, beating several popular 27B-35B models.
- Ran on 2x RTX 3060 12GB with 64GB RAM, stock llama.cpp and Hermes, no gimmicks
- The platform takes 10-45 min per run and has seen hundreds of models submitted by users
- The author urges people to test more 4B-12B models instead of the usual Qwen variants
More from Models
- Gemini App says it was optimized for the new iPhone Duo on day one — GeminiApp · 2026-10-31
- Second dev confirms GPT-6.1 bloat woes, scrapping v1 to rebuild from first principles — Yamapama · 2026-10-31
- Debating ExploitBench: would 96.8% from a flash model break the nonlinear-intelligence thesis? — teortaxesTex · 2026-10-31
- Nathan Lambert: the open-weights cyber weapon scare hasn't materialized—GLM-5.3 proves the point — natolambert · 2026-10-31
- Microsoft releases 4B hybrid computer-use agent scoring 66% on OSWorld — _akhaliq · 2026-10-31
- Claude product lead explains Max-to-Ultra rename and 5-hour limit removal — edwinarbus · 2026-10-31