Ling 3 Tiny Runs at 36 tok/s on 4GB VRAM, Matching Qwen3.5 9B

Benchmarked by Artificial Analysis to outperform Qwen3.5 9B on reasoning, Ling 3.0 Tiny (8B total, 1.3B active params) achieved about 36 tokens/s on a 4GB VRAM machine, notably faster than Qwen 3.5 9B on the same hardware.

2026-08-18 ~ 2026-08-18 · 2 related posts

• Ling 3.0 Tiny hits 36 tok/s on 4GB VRAM, matching Qwen 3.5 9B in quality — cosmos_hu · 2026-08-18
• Is Ling 3 Tiny underrated? Benchmarks suggest it beats Qwen3.5 9B — Hot_Example_4456 · 2026-08-18