Qwen3.8-27B rivals frontier models, runs on a single RTX 3090

yaboyyoungairvent · reddit · 2026-08-18

Qwen3.8-27B has achieved impressive results on the Artificial Analysis Benchmark, landing next to DeepSeek V4 and GPT-5.6 Luna Max. This allows users to run a near-frontier model using just a single RTX 3090 GPU, significantly lowering the hardware barrier for local high-performance LLMs.

Related event: Qwen3.8-27B matches top closed-source models on benchmarks while running on a single consumer GPU(20 posts)→

Original post →

More from Models

• User preference: Sticking with Sonnet over Opus for coding — timfreemints · 2026-08-24
• Local DeepSeek V4 Flash Benchmark: 24 tok/s on Epyc + RTX 5090 — IntravenusDeMilo · 2026-08-24
• Speculation: Ox Alpha Allegedly a Multi-Party Collaboration — teortaxesTex · 2026-08-24
• Carnice-V3-27b: Beats 10x Larger Models Locally — TheMoonMidas · 2026-08-24
• Users Notice ChatGPT Becomes Less Agreeable, Stands Firm in Debates — USSGoat · 2026-08-24
• Grok ports DOOM to ESP32-S chip at high frame rate — yunta_tsai · 2026-08-24

Models channel →