Grouped by reported model identity. Exact weights, quantizations and software can differ. How results are reviewed
Documented results
24 selected tests| View ReportCustom evaluation | Qwen3-14B | Local LLM custom suite4b · Math / math_basic | 100 %Score · 1 / 1 · Custom answer parser 52.22 tok/sThroughput · per request | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View ReportCustom evaluation | Qwen3-14B | Custom C++ corpus perplexity~960KB C++ corpus2K configured context | 1.7031 perplexityScore · Perplexity · Lower is better | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | 1,177.22 tok/sPrefill · mean · per request | — | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | — | 8.1 tok/sGeneration · per request | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 4,500.71 tok/sTotal throughput · mean · aggregate | — | — | 2 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14BConflicting evidence | Lemonade benchmark2026.39.1 · code-debug | 65.5 tok/sThroughput · mean · unknown | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 53.73 tok/sThroughput · unknown | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Relative performance | 1.39 x vs RTX 4500 BlackwellRelative performance per dollar · Whole system | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Local LLM custom suite4b · Math / math_discount | 100 %Score · 1 / 1 · Custom answer parser 52.73 tok/sThroughput · per request | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Custom C++ corpus perplexity~960KB C++ corpus2K configured context | 1.7044 perplexityScore · Perplexity · Lower is better | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | — | 56.38 tok/sGeneration · mean · per request | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | 274 tok/sPrefill · per request | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 2,956.34 tok/sTotal throughput · mean · aggregate | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14BConflicting evidence | Lemonade benchmark2026.39.1 · code-debug | 61.7 tok/sThroughput · mean · unknown | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 63.52 tok/sThroughput · unknown | — | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Local LLM custom suite4b · Math / math_speed | 100 %Score · 1 / 1 · Custom answer parser 52.87 tok/sThroughput · per request | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Custom C++ corpus perplexity~960KB C++ corpus2K configured context | 1.7134 perplexityScore · Perplexity · Lower is better | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | 1,104 tok/sPrefill · per request | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 511.75 tok/sThroughput · mean · aggregate | — | — | 2 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14BConflicting evidence | Lemonade benchmark2026.39.1 · code-short | 64.9 tok/sThroughput · mean · unknown | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Local LLM custom suite4b · Math / math_probability | 100 %Score · 1 / 1 · Custom answer parser 52.87 tok/sThroughput · per request | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-14B | Custom C++ corpus perplexity~960KB C++ corpus2K configured context | 1.7544 perplexityScore · Perplexity · Lower is better | — | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-14B | Token speed | — | — | 54.3 tok/sGeneration · per request | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-14B | Token speed | 336.15 tok/sThroughput · mean · aggregate | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
Quantizations and software
Quantizations: FP8 dynamic, FP8-dynamic, Q4_K_M
Engines: Lemonade / llama.cpp, llama.cpp, llama.cpp TurboQuant fork, Ollama, vLLM
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: R9700 · RX 9070 XT · RX 9070 · RX 7900 XTX