RTX 3090 x2
10recorded runs
240.8best tok/s
7sourced
How it compares — 24 GB per card · multi-GPU
What fits on 24 GB — measured
| Model | Quant | Peak VRAM | Headroom |
|---|---|---|---|
| Qwen3.8-27B | GPTQ-4bit | 45.3 GB | -21.3 GB |
| Qwen3.8-Flash-Next | Unsloth-Dynamic-IQ3_XXS | 42.9 GB | -18.9 GB |
| Qwen3.8-Flash-Next-REAP-256-duo | Unsloth-Dynamic-Q3_K_XL | 33.9 GB | -9.9 GB |
| NVIDIA-Nemotron-3.5-Lightning-30B-A3B | NVFP4 | 19.5 GB | 4.5 GB |
All benchmarks Compare RTX 3090 x2 →
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| RTX 3090 x2 · 24 GB | Qwen3.6-27B | AutoRound INT4 | vllm 0.19.1 | 200000 | 240.8 ✓ | — | — | src |
| RTX 3090 x2 · 24 GB | Qwen3.8-27B | AutoRound INT4 + fp8 KV | sglang 0.5.19 | 262000 | 219.8 ✓ | — | — | src |
| RTX 3090 x2 · 24 GB | NVIDIA-Nemotron-3.5-Lightning-30B-A3B | NVFP4 | vllm 0.27.1 | 131072 | 197.0 ✓ | 6719.0 | 19.5 | src |
| RTX 3090 x2 · 24 GB | Qwen3.8-27B | W4A16-AutoRound | vllm 0.27.1+syv-ai.00210159 | 262144 | 182.4 ✓ | 1338.2 | — | src |
| RTX 3090 x2 · 24 GB | MiniCPM-V 4.7-35B-A3B | int4 | vllm 0.31.1rc1.dev296 (nightly fe4df3d) | 262144 | 172.2 ✓ | — | — | src |
| RTX 3090 x2 · 24 GB | Qwen3.8-Flash-Next-REAP-256-duo | Unsloth-Dynamic-Q3_K_XL | llama.cpp 0.3.0-dev / commit 035e227 (llamacpp… | 32768 | 53.5 ✓ | 948.6 | 33.9 | src |
| RTX 3090 x2 · 24 GB | Qwen3.8-Flash-Next | Unsloth-Dynamic-IQ3_XXS | llama.cpp 0.3.0-dev / commit 035e227 (llamacpp… | 217088 | 40.1 ✓ | 340.1 | 42.9 | src |
| RTX 3090 x2 · 24 GB | Ling-3.0-tiny | Q8_0 | llama.cpp — | 2048 | 232.8 … | 9243.2 | — | src |
| RTX 3090 x2 · 24 GB | Qwen3.8-27B | GPTQ-4bit | vllm — | 8192 | 96.6 … | 1311.1 | 45.3 | src |
| RTX 3090 x2 · 24 GB | Muse-Glimmer-30B | Unsloth-Dynamic-Q6_K_XL | llama.cpp — | 2048 | 31.7 … | 1604.0 | — | src |