RTX 4070
13recorded runs
64.4best tok/s
0sourced
How it compares — 12 GB tier
What fits on 12 GB — measured
| Model | Quant | Peak VRAM | Headroom |
|---|---|---|---|
| BTL-4-Compact | IQ2_XXS | 11.9 GB | 0.1 GB |
| Qwen3.6-27B | UD-IQ2_M | 11.8 GB | 0.2 GB |
| Ornith-1.5-35B-A3B-MTP | Q4_K_M | 11.8 GB | 0.2 GB |
| Ornith-1.5-35B-A3B | Q6_K | 11.6 GB | 0.4 GB |
| Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-YMQ | XS-Pro | 11.5 GB | 0.5 GB |
| badtheorylabs_BTL-4 | Q4_K_L | 11.5 GB | 0.5 GB |
| Qwen3.8-27B | GSQ-RCO-IQ2_XS | 11.5 GB | 0.5 GB |
| Qwen3.6-27B-i1 | IQ1_S | 11.4 GB | 0.6 GB |
All benchmarks Compare RTX 4070 →
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| RTX 4070 · 12 GB | Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-YMQ | XS-Pro | llama.cpp — | 20480 | 64.4 … | 29.4 | 11.5 | src |
| RTX 4070 · 12 GB | BTL-4-Compact | IQ2_XXS | llama.cpp — | 131072 | 58.5 … | 1336.6 | 11.9 | src |
| RTX 4070 · 12 GB | Ornith-1.5-35B-A3B-MTP | Q4_K_M | llama.cpp — | 102400 | 55.3 … | 585.6 | 11.8 | src |
| RTX 4070 · 12 GB | badtheorylabs_BTL-4 | Q4_K_L | llama.cpp — | 180000 | 55.1 … | 38.1 | 11.5 | src |
| RTX 4070 · 12 GB | Ornith-1.5-35B-A3B | Q6_K | llama.cpp — | 215000 | 42.7 … | 36.0 | 11.6 | src |
| RTX 4070 · 12 GB | Kwaipilot_KAT-Coder-V2.5-Dev | Q4_K_L | llama.cpp — | 262144 | 42.2 … | 64.4 | 4.4 | src |
| RTX 4070 · 12 GB | Qwen3.8-27B | UD-IQ3_XXS | llama.cpp — | 32768 | 36.5 … | 65.2 | 10.8 | src |
| RTX 4070 · 12 GB | NVIDIA-Nemotron-3.5-Lightning-30B-A3B | IQ4_XS | llama.cpp — | 262144 | 35.0 … | 398.2 | 11.3 | src |
| RTX 4070 · 12 GB | Qwen3.6-27B | UD-IQ2_M | llama.cpp — | 65536 | 23.5 … | 456.8 | 11.8 | src |
| RTX 4070 · 12 GB | Qwen3.8-27B | GSQ-RCO-IQ2_XS | llama.cpp — | 152000 | 17.6 … | 315.0 | 11.5 | src |
| RTX 4070 · 12 GB | Qwen3.6-27B-i1 | IQ1_S | llama.cpp — | 196608 | 15.5 … | 220.1 | 11.4 | src |
| RTX 4070 · 12 GB | Qwen3.8-27B-Ridge | Ridge-3.7bpw | llama.cpp — | 65536 | 9.0 … | 12.8 | 11.0 | src |
| RTX 4070 · 12 GB | Godoter-27B | Q4_K_M | llama.cpp — | 32768 | 3.2 … | 446.5 | 9.5 | src |