RTX 3080
10recorded runs
277.8best tok/s
9sourced
What fits on 10 GB — measured
| Model | Quant | Peak VRAM | Headroom |
|---|---|---|---|
| Qwen3.8-27B | IQ2_XXS | 9.7 GB | 0.3 GB |
| gemma-4-12B-it-qat | Q4_0 | 9.7 GB | 0.3 GB |
| Ornith-1.5-35B-A3B | Q4_K_M | 9.6 GB | 0.4 GB |
| Qwen3.8-27B-Uncensored | IQ2_XXS | 9.5 GB | 0.5 GB |
| gpt-oss-20b | MXFP4 | 9.3 GB | 0.7 GB |
| Muse-Glimmer-30B | Q4_K_M | 8.8 GB | 1.2 GB |
| Qwen3-14B | Q4_K_M | 8.7 GB | 1.3 GB |
| Ornith-1.5-9B | Q6_K | 8.5 GB | 1.5 GB |
All benchmarks Compare RTX 3080 →
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| RTX 3080 · 10 GB | LFM2.5-8B-A1B | Q6_K | llama.cpp b10472 | 128000 | 277.8 ✓ | 9840.7 | 8.1 | src |
| RTX 3080 · 10 GB | Ornith-1.5-9B | Q6_K | llama.cpp b10472 | 131072 | 82.8 ✓ | 3397.4 | 8.5 | src |
| RTX 3080 · 10 GB | Qwen3.8-27B-Uncensored | IQ2_XXS | llama.cpp b10472 | 16384 | 74.1 ✓ | 1159.2 | 9.5 | src |
| RTX 3080 · 10 GB | Bonsai-27B | Q1_0 | llama.cpp b10472 | 32768 | 72.9 ✓ | 1319.8 | 5.1 | src |
| RTX 3080 · 10 GB | gemma-4-12B-it-qat | Q4_0 | llama.cpp b10472 | 262144 | 71.2 ✓ | 2671.1 | 9.7 | src |
| RTX 3080 · 10 GB | Ornith-1.5-35B-A3B | Q4_K_M | llama.cpp b10472 | 131072 | 64.4 ✓ | 596.3 | 9.6 | src |
| RTX 3080 · 10 GB | gpt-oss-20b | MXFP4 | llama.cpp b10472 | 32768 | 57.1 ✓ | 1463.0 | 9.3 | src |
| RTX 3080 · 10 GB | Qwen3.8-27B | IQ2_XXS | llama.cpp b10472 | 65536 | 44.4 ✓ | 1079.4 | 9.7 | src |
| RTX 3080 · 10 GB | Muse-Glimmer-30B | Q4_K_M | llama.cpp b10472 | 16384 | 6.0 ✓ | 672.4 | 8.8 | src |
| RTX 3080 · 10 GB | Qwen3-14B | Q4_K_M | llama.cpp — | 2048 | 75.0 … | — | 8.7 | src |