Carnice-V3-27b: Beats 10x Larger Models Locally
TheMoonMidas · x · 2026-08-24
- Release: Carnice-V3-27b, a fully open-source 27B parameter model built on Qwen2.5-27B.
- Performance: Claims to beat models 10x its size in Hermes-agent benchmarks.
- Hardware: Optimized for consumer GPUs (3090+) for local inference.
- Availability: BF16 weights and GGUF formats are available for download.
More from Models
- China's StepFun releases Step 5 Preview weights: 600B MoE, 27B active, 1M context — Dr_Singularity · 2026-10-15
- StepFun open-sources Step 5 Preview: 600B MoE with 1M context, No. 2 on OpenRouter — Dr_Singularity · 2026-10-15
- Google says teams across the company are using Gemini 4 Argon on its hardest engineering problems — CurieuxExplorer · 2026-10-15
- Perplexity's new embedding models: index with a 9B model, search with 0.6B — antoine_chaffin · 2026-10-15
- 797GB of n-gram counts in CPU RAM let 1B models match 57% more compute — PangWeiKoh · 2026-10-15
- Leak: Moonshot's Kimi flagship 'x4' done with pre-training, K3.1 in internal testing — lyraxana · 2026-10-15