AirLLM update runs 2.8T Kimi K3 on under 4GB VRAM
pmttyji · reddit · 2026-08-20
AirLLM released updates significantly reducing inference memory, enabling massive models on consumer hardware. It now supports Qwen3.8-27B (3.33GB VRAM) and the 2.8T parameter Kimi K3 (3.72GB VRAM), and can run DeepSeek-V3 (671B) on 12GB. The tool uses per-expert streaming for MoE models, requiring no quantization, distillation, or pruning.
More from Infra
- Prediction: Opus 5.5-level models on laptops by 2027, backed by Qwen 27B data — cephaloform · 2026-10-16
- Oracle down 12% in a week, nearly 50% off June high as OpenAI exposure spooks market — alvelda · 2026-10-16
- US data center moratoriums top 400 — an AI bull explains why they're actually good for AI — kevinsxu · 2026-10-16
- Microsoft Research unveils Magnetar, the fastest zkVM at 10-39x speedups and post-quantum security — jedisct1 · 2026-10-16
- Bonsai 2 2B hits 120 tok/s on iPhone 18 Pro, outpaces Qwen 3.5 2B at half the memory — cephaloform · 2026-10-16
- Samsung cuts Q4 phone output as Nittobo shifts E-glass to AI-chip T-glass — zephyr_z9 · 2026-10-16