Qwen3-Coder-30B-A3B-Instruct on AMD GPUs
404 results · 9 sources · 526 measurements
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-Coder-30B-A3B-InstructConflicting evidence | HumanEval100 problems | 98 %Reported pass@1 · n = 100 | — | 94.08 tok/sGeneration · mean · per request | 2 × W7800Layer split · 48 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View ReportCustom evaluation | Qwen3-Coder-30B-A3B-InstructConflicting evidence | HumanEval100 problems | 99 %Reported pass@1 · n = 100 | — | 96.97 tok/sGeneration · mean · per request | 1 × W7800Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 3,064.51 tok/sPrefill · unknown | 117.27 tok/sGeneration · unknown | 2 × RX 7900 XTXLayer split · 24 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 239.5 tok/sPrefill · unknown | 14.4 tok/sGeneration · unknown | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 2,454.73 tok/sPrefill · unknown | 82.03 tok/sGeneration · peak · unknown | 3 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-Coder-30B-A3B-Instruct | Token speed512 input tokens · 0 existing KV tokens | — | 321.56 tok/sPrefill · mean · per request | — | 2 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportServing benchmark | Qwen3-Coder-30B-A3B-Instruct | Token speed16 coding prompts repeated twice | 104.42 tok/sThroughput · aggregate | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 974 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speed | — | 2,104 tok/sPrefill · unknown | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speedsynth-bulk-dispatcher-impl | — | — | 93.55 tok/sGeneration · mean · per request | 2 × W7800Layer split · 48 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speedsynth-bulk-dispatcher-impl | — | — | 95.77 tok/sGeneration · mean · per request | 1 × W7800Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 2,532.29 tok/sPrefill · unknown | 136.45 tok/sGeneration · unknown | 2 × RX 7900 XTXLayer split · 24 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 736.1 tok/sPrefill · unknown | 39.6 tok/sGeneration · unknown | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 2,766.5 tok/sPrefill · unknown | 75.92 tok/sGeneration · peak · unknown | 3 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-Coder-30B-A3B-Instruct | Token speed128 output tokens · 0 existing KV tokens | — | — | 18.51 tok/sGeneration · mean · per request | 2 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View ReportServing benchmark | Qwen3-Coder-30B-A3B-Instruct | Token speed16 coding prompts repeated twice | 126.68 tok/sThroughput · aggregate | — | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | — | 113.6 tok/sGeneration · mean · per request | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speed | — | — | 113 tok/sGeneration · unknown | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speedsynth-horner-mpmath-derivative | — | — | 94.63 tok/sGeneration · mean · per request | 2 × W7800Layer split · 48 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3-Coder-30B-A3B-InstructConflicting evidence | Token speedsynth-horner-mpmath-derivative | — | — | 97.39 tok/sGeneration · mean · per request | 1 × W7800Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 230 tok/sPrefill · unknown | 43.8 tok/sGeneration · unknown | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | 429.2 tok/sPrefill · unknown | 24.2 tok/sGeneration · unknown | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportLocalMaxxing | Qwen3-Coder-30B-A3B-Instruct | Token speed | — | — | 87.15 tok/sGeneration · peak · unknown | 3 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3-Coder-30B-A3B-Instruct | Token speed2,048 input tokens · 16,384 existing KV tokens | — | 198 tok/sPrefill · mean · per request | — | 2 × R9700Multi-GPU · split unknown · 32 GB VRAM/GPUConfiguration
|
Quantizations and software
Quantizations: AWQ INT4 G32, BF16, Q4_K_M, Q4_K_XL, Q8_0, UD-Q8_K_XL
Engines: llama.cpp, Ollama, Paiton + vLLM, vLLM
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: W7800 · R9700 · RX 7900 XTX · Ryzen AI Max+ 395