RX 7900 XTX LLM benchmarks
909 results · 2,179 measurements · 37 sources
Device specifications
| Variant | Memory | Memory type | Bus | Bandwidth | FP16 matrix | INT8 matrix |
|---|---|---|---|---|---|---|
| RX 7900 XTX ↗ | 24 GB | GDDR6 | 384-bit | 960 GB/s | 123 TFLOPS | 123 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 909 resultsSource-reported results, selected across reports by documentation quality. Match configurations before comparing values.
| View ReportGeneric LLM | Llama-3-8B-Instruct | LoRA trainingyahma/alpaca-cleaned; packed; revision unreported | 1.9062 lossFinal-step training loss · Cross-entropy on the final training step; not held-out evaluation · Lower is better 1,073.57 tok/sTraining · per GPU | — | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
|---|---|---|---|---|---|---|
| View Reportllm-bench.io | Qwen3.8-27BConflicting evidence | llm-bench.io QualityFour sequential scenarios: agent workflow, code generation, role play, research | 87.88 / 100LLM judge · Mean | 547.47 tok/sPrefill · mean · per request | 64.03 tok/sGeneration · mean · per request | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportCustom evaluation | Qwen3.8-Flash-Next | AIME24LLST protocol 1.0 subset · 10 pinned LLST v1.0 questions; not the full AIME24 benchmark32K configured context | 30 %Score · n = 100 · Strict numeric extraction | — | — | 2 × RX 7900 XTXMulti-GPU · split unknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.8-27B | Token speed | — | 1,016.56 tok/sPrefill · per request | 27.55 tok/sGeneration · per request | Mixed: RX7900XTX24GB + RX7700XT12GB · 2 GPUsPipeline parallelConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 5,776.99 tok/sPrefill · mean · unknown | 172.28 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.1-8B-Instruct | Token speed | — | 1,044.41 tok/sPrefill · mean · unknown | 60.37 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen2.5-14B-Instruct | Token speed | — | 591.81 tok/sPrefill · mean · unknown | 37.7 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.1-8B-Instruct | Token speed | — | 1,068.59 tok/sPrefill · mean · unknown | 7.71 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen2.5-14B-Instruct | Token speed | — | 592.73 tok/sPrefill · mean · unknown | 4.16 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Llama-3.2-1B-Instruct | Token speed | — | 5,076.32 tok/sPrefill · mean · unknown | 16.43 tok/sGeneration · mean · unknown | GPU count unknownUnknown · 24 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 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 Reportllama-bench | DeepSeek-R1-Distill-Llama-8B | Token speed | — | 2,970.26 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.8-27B | Token speedLong output gen512 | — | — | 97.38 tok/sGeneration · per request | 2 × RX 7900 XTXTensor parallel · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Qwen3.6-27B-Fable-Fus-711-UnHeretic-NM-DAU-NEO-MAX-NEO-LOW-MTP | Token speed | — | 1,075.8 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 758.55 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View ReportGeneric LLM | Qwen3.8-27B | Token speed | ~40.4 tok/sThroughput · per request | — | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 3,243.15 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 2,817.81 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 3,434.01 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 3,386.75 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 3,473.24 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | gemma-4-31B-it | Token speed | — | 920.55 tok/sPrefill · mean · per request | — | 2 × RX 7900 XTXLayer split · 24 GB VRAM/GPUConfiguration
|
| View Reportllama-bench | Llama-2-7B | Token speed | — | 4,145.55 tok/sPrefill · mean · per request | — | 1 × RX 7900 XTXSingle GPU · 24 GB VRAM/GPUConfiguration
|
Models and software in this collection
Model families: DeepSeek Coder V2, DeepSeek R1, Devstral Small 2, GLM 4.7, GPT-OSS, Gemma 4, Granite 3.0, Laguna, Llama 2, Llama 3, Llama 3.1, Llama 3.2, Ministral 3, Mistral, Mistral Small 3.2, Muse Glimmer, Qwen2.5, Qwen25, Qwen3, Qwen3.5, Qwen3.6, Qwen3.8, Qwen38
Engines: LM Studio, Ollama, SGLang, Strata, Unsloth Studio, hipfire, llama.cpp, llamafile, lmstudio, qwen-kernel, torchtune, vLLM
Methodology and evidence standards · Download the research dataset