llama3.1-70b vs qwen2.5-14b on RTX 3080
Compare llama3.1-70b and qwen2.5-14b running locally on RTX 3080 (10GB VRAM) — tokens/sec, VRAM fit, and quality scores for local inference.
Compare a different combination
| Quant | Q4_K_M |
| VRAM required | 42GB |
| Fits on RTX 3080 | ✗ No |
| Quality score | 66 / 100 |
| Eval sources | MMLU · MMLU-Pro · IFEval · GSM8K · ARC-C · HumanEval · MBPP+ |
| Data | estimate |
| Quant | Q4_K_M |
| VRAM required | 9GB |
| Fits on RTX 3080 | Tight |
| Quality score | 61 / 100 |
| Eval sources | MMLU-Pro · IFEval · BBH · LiveCodeBench |
| Data | estimate |
Add a source link or correction so this page can rank with real hardware-specific data instead of estimates.
Contribute GPU comparison dataGPU comparison FAQ
Which is faster on RTX 3080: llama3.1-70b or qwen2.5-14b?
This RTX 3080 comparison shows VRAM fit and quality data for llama3.1-70b vs qwen2.5-14b, but one or both decode-speed benchmarks are still missing.
Do both models fit on RTX 3080?
llama3.1-70b: ✗ No. qwen2.5-14b: Tight. The fit labels use Q4_K_M VRAM estimates and the benchmark data available for this GPU.
How much VRAM do these models need?
llama3.1-70b needs about 42GB in Q4_K_M; qwen2.5-14b needs about 9GB in Q4_K_M.
Why can speed differ by GPU?
Local LLM speed depends on GPU memory bandwidth, backend, quantization, context length, and whether the model fits entirely in VRAM or unified memory.
How can I correct this comparison?
Use the GitHub correction link on this page with your GPU, model, quantization, context length, and measured decode tokens per second.