LocalLLMLeaderboard
Home / By Size / 27b

Best 27b Models in 2026 (Updated Weekly)

As of 2026-W40, the best 27b model is undefined (Intelligence undefined).
#ModelParamsIntelligenceCodingDownloads 30d
1Qwen3.8 27B27.8B——7.02M
2Qwen3.8 27B TURBO Fable Cold Fusion 735 882 Heretic Uncensored NEO CODER MAX MTP GGUF26.9B——1.73M
3Swift 1.5 Qwen3.8 27B GSQ RCO GGUF26.9B——0.12M
4MiniMax H3 TextEncoder Qwen3VL 32B abliterated GGUF26.5B——0.03M
5OrcaSAQ 2 Cyber 27B Uncensored GGUF27.3B——0.01M