Qwen3.8-27B-GSQ-RCO on AMD GPUs

12 results · 2 sources · 48 measurements

Grouped by reported model identity. Exact weights, quantizations and software can differ. How results are reviewed

Documented results

12 selected tests


View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	89.82 / 100LLM judge · Mean	332.77 tok/sPrefill · mean · per request	49.83 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp k8 v4 -1 budget Backend Not reported Harness llm-benchmark 0.4.89+102 Prompt / output ? / ? tokens Context 128000 Concurrency 1 Date 2026-10-09 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SWeights ↗	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	85.82 / 100LLM judge · Mean	205.79 tok/sPrefill · mean · per request	29.91 tok/sGeneration · mean · per request	GPU count unknownUnknown · 24 GB VRAM/GPU Configuration Memory 24 GB VRAM/GPU OS Windows 11 Home 25H2 Engine Ollama Backend Not reported Harness llm-benchmark 0.4.59+97 Prompt / output ? / ? tokens Context 160000 Concurrency 1 Date 2026-09-17 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	88.76 / 100LLM judge · Mean	335.62 tok/sPrefill · mean · per request	50.6 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp q8 q4 -1 budget Backend Not reported Harness llm-benchmark 0.4.89+102 Prompt / output ? / ? tokens Context 128000 Concurrency 1 Date 2026-10-09 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SWeights ↗	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	85.55 / 100LLM judge · Mean	732.9 tok/sPrefill · mean · per request	34.37 tok/sGeneration · mean · per request	GPU count unknownUnknown · 24 GB VRAM/GPU Configuration Memory 24 GB VRAM/GPU OS Windows 11 Home 25H2 Engine Ollama Backend Not reported Harness llm-benchmark 0.4.82+98 Prompt / output ? / ? tokens Context 65536 Concurrency 1 Date 2026-09-27 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	85.46 / 100LLM judge · Mean	261.33 tok/sPrefill · mean · per request	38.66 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp k8 v8 -1 budget Backend Not reported Harness llm-benchmark 0.4.89+102 Prompt / output ? / ? tokens Context 65536 Concurrency 1 Date 2026-10-09 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SWeights ↗	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	84.65 / 100LLM judge · Mean	192.91 tok/sPrefill · mean · per request	18.57 tok/sGeneration · mean · per request	GPU count unknownUnknown · 24 GB VRAM/GPU Configuration Memory 24 GB VRAM/GPU OS Windows 11 Home 25H2 Engine Ollama Backend Not reported Harness llm-benchmark 0.4.71+98 Prompt / output ? / ? tokens Context 180000 Concurrency 1 Date 2026-09-20 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	84.4 / 100LLM judge · Mean	333.13 tok/sPrefill · mean · per request	50.21 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp q5 q5 -1 budget Backend Not reported Harness llm-benchmark 0.4.89+102 Prompt / output ? / ? tokens Context 128000 Concurrency 1 Date 2026-10-09 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	83.39 / 100LLM judge · Mean	328.58 tok/sPrefill · mean · per request	48.97 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp k8 v4 -1 budget Backend Not reported Harness llm-benchmark 0.4.89+102 Prompt / output ? / ? tokens Context 128000 Concurrency 1 Date 2026-10-09 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	86.71 / 100LLM judge · Mean	306.15 tok/sPrefill · mean · per request	40.72 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp Backend Not reported Harness llm-benchmark 0.4.87+102 Prompt / output ? / ? tokens Context 65536 Concurrency 1 Date 2026-10-04 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	86.19 / 100LLM judge · Mean	312.66 tok/sPrefill · mean · per request	34.14 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp Backend Not reported Harness llm-benchmark 0.4.87+102 Prompt / output ? / ? tokens Context 65536 Concurrency 1 Date 2026-10-04 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	83.83 / 100LLM judge · Mean	304 tok/sPrefill · mean · per request	33.79 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp Backend Not reported Harness llm-benchmark 0.4.87+102 Prompt / output ? / ? tokens Context 65536 Concurrency 1 Date 2026-10-06 Full record →
View Reportllm-bench.io	Qwen3.8-27B-GSQ-RCOIQ3_SRepository unresolved	llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research	80.07 / 100LLM judge · Mean	338.69 tok/sPrefill · mean · per request	33.3 tok/sGeneration · mean · per request	GPU count unknownUnknown · 16 GB VRAM/GPU Configuration Memory 16 GB VRAM/GPU OS CachyOS Engine llama.cpp Backend Not reported Harness llm-benchmark 0.4.87+102 Prompt / output ? / ? tokens Context 196608 Concurrency 1 Date 2026-10-04 Full record →

Quantizations and software

Quantizations: IQ3_S
Engines: llama.cpp, llama.cpp k8 v4 -1 budget, llama.cpp k8 v8 -1 budget, llama.cpp q5 q5 -1 budget, llama.cpp q8 q4 -1 budget, Ollama

Explore all 12 results →

This is a selection across sources, not a ranking or a matched experiment. Reading benchmark reports

GPU benchmark pages: RX 7900 XTX · RX 7800 XT