RX 7800 XT LLM benchmarks
228 results · 551 measurements · 14 sources
Device specifications
| Variant | Memory | Memory type | Bus | Bandwidth | FP16 matrix | INT8 matrix |
|---|---|---|---|---|---|---|
| RX 7800 XT ↗ | 16 GB | GDDR6 | 256-bit | 624 GB/s | 74.6 TFLOPS | 74.6 TOPS |
Manufacturer specifications. Shared system RAM is not dedicated VRAM. Unreported values remain —; measured LLM results retain their reported configuration.
Selected reported tests
24 of 228 resultsSource-reported results, selected across reports by documentation quality. Match configurations before comparing values.
| View Reportllm-bench.io | Darwin-28B-REASON.i1 | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 71 / 100LLM judge · Mean | 256.2 tok/sPrefill · mean · per request | 21.11 tok/sGeneration · mean · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View ReportLocalMaxxing | Qwen3.8-27B | Token speed | — | 500 tok/sPrefill · unknown | 42 tok/sGeneration · unknown | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 3,787.39 tok/sPrefill · mean · unknown | 138.8 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 3,395.56 tok/sPrefill · mean · unknown | 138.01 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 3,885.2 tok/sPrefill · mean · unknown | 129.53 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | granite-3.1-8b-instruct | Token speed | — | 581.67 tok/sPrefill · mean · unknown | 37.01 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| 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 Reportllama-bench | llama 8B | Token speed512 input tokens · 0 output tokens · 8,192 existing KV tokens | — | 763.53 tok/sPrefill · mean · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | DeepSeek-R1-Distill-Llama-8B | Token speed | — | 1,811.3 tok/sPrefill · mean · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Ternary-Bonsai-27B | Token speed | — | 523.14 tok/sPrefill · mean · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.8-27B | Token speed | — | 468.35 tok/sPrefill · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 2,151.81 tok/sPrefill · mean · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 2,109.38 tok/sPrefill · mean · per request | — | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 2,017.33 tok/sPrefill · mean · per request | — | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View Reportllm-bench.io | Darwin-28B-REASON.i1 | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 73.1 / 100LLM judge · Mean | 357.17 tok/sPrefill · mean · per request | 38.41 tok/sGeneration · mean · per request | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportLocalMaxxing | Qwen3.8-27B | Token speed | — | 500 tok/sPrefill · unknown | 28 tok/sGeneration · unknown | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 2,585.45 tok/sPrefill · mean · unknown | 29.68 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 2,579.64 tok/sPrefill · mean · unknown | 29.62 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 2,563.97 tok/sPrefill · mean · unknown | 28.98 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | granite-3.1-8b-instruct | Token speed | — | 435.24 tok/sPrefill · mean · unknown | 6.76 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 16 GB VRAM/GPUConfiguration
|
| View ReportServing benchmark | Qwen3.6-35B-A3B | Token speed | — | 114.5 tok/sPrefill · per request | 57.8 tok/sGeneration · per request | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | llama 8B | Token speed0 input tokens · 128 output tokens · 8,192 existing KV tokens | — | — | 74.81 tok/sGeneration · mean · per request | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | DeepSeek-R1-Distill-Llama-8B | Token speed | — | — | 66 tok/sGeneration · mean · per request | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Ternary-Bonsai-27B | Token speed | — | — | 47.2 tok/sGeneration · mean · per request | 1 × RX 7800 XTSingle GPU · 16 GB VRAM/GPUConfiguration
|
Models and software in this collection
Model families: DeepSeek R1, GLM 4.7, GPT-OSS, Gemma 4, Granite 3.0, Granite 3.1, Llama 2, Llama 3.1, Llama 3.2, Mistral, Qwen2.5, Qwen3, Qwen3.5, Qwen3.6, Qwen3.8
Engines: Ollama, Strata, llama.cpp, llama.cpp (Prism fork), llama.cpp k8 v4, llama.cpp k8 v4 -1 budget, llama.cpp k8 v8 -1 budget, llama.cpp q5 q5 -1 budget, llama.cpp q8 q4 -1 budget, llama.cpp q8 q8, llama.cpp q8 q8 -1 budget, llamafile, vLLM
Methodology and evidence standards · Download the research dataset