Can Yi 34B 200K run on the RTX 3060 12GB?
Not fully in VRAM. Even at 2-bit with 8K context, Yi 34B 200K needs about 16.0 GB, more than the 11.4 GB usable on the RTX 3060 12GB. It can still run with offload (slower): Q4_K_M with part of the model in system RAM, about 2.0–3.8 tokens/sec (estimate). Details and command.
Runs with offload (slower)
Runs with offload 4-bit (Q4_K_M) at 8K context: 29 of 60 layers on the RTX 3060 12GB (about 10.7 GB) and the other 31 in system RAM (about 10.9 GB). Needs a PC with at least 32 GB of system RAM. Estimated speed: 2.0–3.8 tokens/sec (slow but usable), versus a bandwidth ceiling of ~17 tok/s if it fit fully in VRAM.
llama-server -hf TheBloke/Yi-34B-200K-GGUF:Q4_K_M -c 8192 -ngl 29
Estimate, not a benchmark. The command downloads the GGUF from Hugging Face. The speed range spans dual-channel DDR4-3200 to DDR5-6000 system RAM; your CPU, RAM speed and llama.cpp version move it. If llama.cpp runs out of GPU memory, lower -ngl by 2-3; if VRAM is left over, raise it for more speed. Recent llama.cpp builds can also pick the split automatically. How we estimate offload.
Memory by quant and context
| Quant | 4K ctx | 8K ctx | 32K ctx | 128K ctx | Max tok/s (est., 8K) |
|---|---|---|---|---|---|
| 2-bit (Q2_K) • | 15.0 GB No | 16.0 GB No | 21.6 GB No | 44.1 GB No | — |
| 3-bit (Q3_K_M) • | 17.0 GB No | 17.9 GB No | 23.5 GB No | 46.0 GB No | — |
| 4-bit (Q4_K_M) • | 20.7 GB No | 21.6 GB No | 27.3 GB No | 49.8 GB No | — |
| 5-bit (Q5_K_M) • | 24.1 GB No | 25.0 GB No | 30.7 GB No | 53.2 GB No | — |
| 6-bit (Q6_K) • | 27.7 GB No | 28.7 GB No | 34.3 GB No | 56.8 GB No | — |
| 8-bit (Q8_0) • | 35.5 GB No | 36.4 GB No | 42.1 GB No | 64.6 GB No | — |
| 16-bit (F16/BF16) | 65.5 GB No | 66.5 GB No | 72.1 GB No | 94.6 GB No | — |
• real GGUF file size from TheBloke/Yi-34B-200K-GGUF. Speed is a bandwidth ceiling (360 GB/s), not a benchmark. Methodology.
Hardware that does fit it
RX 7900 XTRTX 3090RTX 3090 TiRTX 4090RX 7900 XTXMac M2 24GB
What fits instead
- Cheapest listed GPU that runs Yi 34B 200K: RTX 5090 (32 GB, launch MSRP $1,999) at Q5_K_M, fully in VRAM at 8K context.
- Largest popular model that fits the RTX 3060 12GB: KORMo Web 11B (11B) at Q5_K_M, about 9.3 GB at 8K context.
"Runs" and "fits" mean the whole model in GPU memory at 8K context (4-bit or better for the model pick). Prices are US launch MSRPs, not today's street prices. "Popular" means the 60 most-downloaded models we track.
More
Yi 34B 200K on every GPU and Mac · Every model that runs on the RTX 3060 12GB · Other models that fit 12 GB VRAM · Try other settings in the calculator
Guides: Best GPU for local LLMs by budget · Can I run gpt-oss, Llama 70B or Qwen3? · Mac vs PC for local LLMs
Data: Hugging Face Hub (01-ai/Yi-34B-200K), NVIDIA specs. Model data last checked 2026-11-17; refreshed hourly.