Grouped by reported model identity. Exact weights, quantizations and software can differ. How results are reviewed
Documented results
24 selected tests| View ReportGeneric LLM | Qwen3.6-35B-A3B | HumanEvalFull 164 problems | 59.15 %pass@1 · 97 / 164 · Executable tests 38 tok/sThroughput · aggregate | — | — | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View Reportllm-bench.io | Qwen3.6-35B-A3B | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 83.52 / 100LLM judge · Mean | — | 161.16 tok/sGeneration · mean · per request | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View Reportllm-bench.io | Qwen3.6-35B-A3B | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 65.71 / 100LLM judge · Mean | 784.16 tok/sPrefill · mean · per request | 60.81 tok/sGeneration · mean · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportCoding session | Qwen3.6-35B-A3BConflicting evidence | PinchBenchFive-task smoke / sanity,calendar,summary,weather,csv_iris_summary1 concurrent request | 4.58 / 5LLM judge | — | — | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportCoding session | Qwen3.6-35B-A3BPartial evaluation | SWE-bench Verified MiniVerified Mini: 50 tasks (25 Django + 25 Sphinx) | 66 %LLM judge · 33 / 50 56 %Exit-zero rate · 28 / 50 · Container test exit | — | — | 1 × R9700Single GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.6-35B-A3B | HEPAValid complete runs64K configured context | 100 %Score · n = 1 · Deterministic validity gate | — | — | 1 × Ryzen AI Max+ 395Single GPU · Unified memory · 128 GB RAMConfiguration
|
| View ReportKyuz MTP | Qwen3.6-35B-A3B | Token speed | — | 146.44 tok/sPrefill · per request | 74.08 tok/sGeneration · per request | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.6-35B-A3B | Token speed | — | 2,924.49 tok/sPrefill · median · per request | 94.95 tok/sGeneration · median · per request | 1 × W7900Single GPU · 48 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.6-35B-A3B | Token speed | — | 1,467.82 tok/sPrefill · median · per request | 248.2 tok/sGeneration · median · per request | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View ReportServing benchmark | Qwen3.6-35B-A3B | Token speed | — | 72.7 tok/sPrefill · per request | 55.7 tok/sGeneration · per request | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.6-35B-A3B | Token speed | — | 257 tok/sPrefill · per request | 168.1 tok/sGeneration · per request | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.6-35B-A3B | Token speed | — | 41.7 tok/sPrefill · per request | 22.7 tok/sGeneration · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.6-35B-A3B | Token speed | — | 913.46 tok/sPrefill · per request | 214.72 tok/sGeneration · per request | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.6-35B-A3B | Token speed | — | 399 tok/sPrefill · unknown | 47 tok/sGeneration · unknown | 1 × RX 9070Single GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.6-35B-A3B | Token speed | — | 1,469.23 tok/sPrefill · median · per request | 280.4 tok/sGeneration · median · per request | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed2,048 input tokens · 0 output tokens · 16,384 existing KV tokens | — | 657.12 tok/sPrefill · mean · per request | — | 1 × Ryzen AI Max+ 395Single GPU · Unified memory · 128 GB RAMConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed | — | 1,639 tok/sPrefill · per request | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.6-35B-A3B | Token speed | 15.6 tok/sThroughput · per request | — | — | 4 × R9700Tensor parallel · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed8,192 input tokens · 0 existing KV tokens | — | 2,714 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed512 input tokens · 0 output tokens · 0 existing KV tokens | — | 525.94 tok/sPrefill · mean · per request | — | 1 × Ryzen AI Max+ 395Single GPU · Unified memoryConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed | — | 706 tok/sPrefill · per request | — | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed2,048 input tokens · 16,384 existing KV tokens | — | 2,306.14 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.6-35B-A3B | Token speed | — | — | 96.57 tok/sGeneration · unknown | 1 × RX 9070 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-35B-A3B | Token speed | — | 2,653.96 tok/sPrefill · mean · per request | — | 1 × R9700Single GPU · 32 GB VRAM/GPUConfiguration
|
Quantizations and software
Quantizations: 4-bit (format unreported), AWQ-4bit, BF16, FP8, INT4 W4A16, IQ4_NL, IQ4_XS, MQ4, MQ4-AWQ, MQ4R, MXFP4_MOE, Q2_K_XL, Q4_K_M, Q4_K_S, Q4_K_XL, Q6_K, Q8_0, Q8_K_XL, Quark-W8A8-INT8, TQ3_4S, UD-IQ2_XXS, UD-IQ3_S, UD-IQ3_XXS, UD-IQ4_NL, UD-IQ4_XS, UD-Q2_K_XL, UD-Q3_K_M, UD-Q4_K_M, UD-Q4_K_XL, UD-Q5_K_M
Engines: Custom HIP LLM runner, hipEngine, hipfire, llama.cpp, llama.cpp TurboQuant fork, LM Studio, lmstudio, qwen-kernel, Strata, vllm, vLLM, vLLM Radiance
This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports
GPU benchmark pages: W7900 · R9700 · RX 9070 XT · RX 9070 · RX 7900 XTX · RX 7900 XT · RX 7900 GRE · RX 7800 XT · RX 7600 XT · Ryzen AI Max+ 395