Skip to content
Run LLM Locally

GPU · AMD

RX 6800 XT for local LLMs: which models run?

With 16 GB of VRAM and 512 GB/s of bandwidth, the RX 6800 XT runs models up to about gpt-oss-20b entirely on the GPU.

  • VRAM16 GB GDDR6
  • Bandwidth512 GB/s
  • Board power300 W
  • ArchitectureRDNA 2
ShareFacebookXWhatsAppLinkedInEmail

Specifications

VRAM
16 GB
Memory type
GDDR6
Memory bus
256-bit
Bandwidth
512 GB/s
Board power
300 W
Backends
Vulkan
Interface
PCIe 4.0 x16
Released
November 2020
Segment
Consumer
FP16 compute
Not available

Official specData updated 24 days agoSource: AMD specifications

What the RX 6800 XT runs

Popular models on RX 6800 XT + 64 GB RAM · 8K context
gpt-oss-20bMXFP4 · 8K~12.9 GB · 92–171 tok/s estimateExcellent
Gemma 2 9B ITQ8_0 · 8K~12.7 GB · 28–44 tok/s estimateExcellent
Qwen3 8BQ8_0 · 8K~10.5 GB · 33–51 tok/s estimateExcellent
Llama 3.1 8B InstructQ8_0 · 8K~10.1 GB · 34–52 tok/s estimateExcellent
DeepSeek R1 Distill Llama 8BQ8_0 · 8K~10.1 GB · 34–52 tok/s estimateExcellent
Qwen2.5 7B InstructQ8_0 · 8K~9.2 GB · 37–57 tok/s estimateExcellent
Mistral 7B Instruct v0.3Q8_0 · 8K~9.1 GB · 37–58 tok/s estimateExcellent
Gemma 3 4B ITQ8_0 · 8K~5.4 GB · 66–109 tok/s estimateExcellent
Qwen3 4BQ8_0 · 8K~6.1 GB · 59–96 tok/s estimateExcellent
Phi-4 Mini Instruct 3.8BQ8_0 · 8K~5.9 GB · 63–101 tok/s estimateExcellent
Llama 3.2 3B InstructQ8_0 · 8K~5.1 GB · 74–120 tok/s estimateExcellent
Qwen3 1.7BQ8_0 · 8K~3.6 GB · 115–198 tok/s estimateExcellent
Llama 3.2 1B InstructQ8_0 · 8K~2.5 GB · 186–316 tok/s estimateExcellent
Qwen3 0.6BQ8_0 · 8K~2.5 GB · 198–389 tok/s estimateExcellent
Mistral Small 3.2 24B InstructIQ4_XS · 8K~14.6 GB · 23–36 tok/s estimateCompatible
Qwen3 14BQ6_K · 8K~14 GB · 24–37 tok/s estimateCompatible
DeepSeek R1 Distill Qwen 14BQ6_K · 8K~14.2 GB · 24–37 tok/s estimateCompatible
Qwen2.5 14B InstructQ6_K · 8K~14.2 GB · 24–37 tok/s estimateCompatible
Phi-4 14BQ6_K · 8K~14.1 GB · 24–37 tok/s estimateCompatible
Mistral Nemo 12B InstructQ8_0 · 8K~14.8 GB · 23–35 tok/s estimateCompatible
Gemma 3 12B ITQ8_0 · 8K~14.1 GB · 24–37 tok/s estimateCompatible
Qwen3 30B-A3BQ3_K_M · 8K~15.3 GB · 78–176 tok/s estimateCompatible with limits
Gemma 3 27B ITQ3_K_M · 8K~15.3 GB · 22–34 tok/s estimateCompatible with limits
Mixtral 8x7B InstructQ3_K_M · 8K~21.5 GB · 9–18 tok/s estimateCPU/GPU offload
Qwen3 32BQ3_K_M · 8K~18.5 GB · 5–10 tok/s estimateCPU/GPU offload
DeepSeek R1 Distill Qwen 32BQ3_K_M · 8K~18.5 GB · 5–10 tok/s estimateCPU/GPU offload
Qwen2.5 32B InstructQ3_K_M · 8K~18.5 GB · 5–10 tok/s estimateCPU/GPU offload
Qwen2.5 Coder 32B InstructQ3_K_M · 8K~18.5 GB · 5–10 tok/s estimateCPU/GPU offload
QwQ 32BQ3_K_M · 8K~18.5 GB · 5–10 tok/s estimateCPU/GPU offload
Qwen3 Coder 30B-A3B InstructQ3_K_M · 8K~15.9 GB · 54–141 tok/s estimateCPU/GPU offload
Gemma 2 27B ITQ3_K_M · 8K~16.4 GB · 11–20 tok/s estimateCPU/GPU offload
gpt-oss-120bMXFP4 · 8K~63.2 GB · 8.5–17 tok/s estimatePossible but slow
Llama 4 Scout 17B-16E InstructQ3_K_M · 8K~55.6 GB · 3–5.5 tok/s estimatePossible but slow
Qwen2.5 72B InstructQ3_K_M · 8K~38.2 GB · 0.9–1.7 tok/s estimateNot recommended
Llama 3.3 70B InstructQ3_K_M · 8K~36.9 GB · 0.9–1.8 tok/s estimateNot recommended
DeepSeek R1 Distill Llama 70BQ3_K_M · 8K~36.9 GB · 0.9–1.8 tok/s estimateNot recommended
DeepSeek R1 (671B)Q8_0 · 8K~699 GB · No dataIncompatible
Qwen3 235B-A22BQ8_0 · 8K~247 GB · No dataIncompatible

For each model, the highest-precision quantization that runs entirely in fast memory; otherwise the one with the best result. Speeds are ranges; "estimate" unless a benchmark exists for this exact hardware.

EstimateData updated 24 days agoSource: Run LLM Locally engine

Limits in plain words

16 GB of VRAM holds the weights and KV cache of models up to about gpt-oss-20b at Q4_K_M. Bigger models spill into system RAM and slow down sharply: generation speed is then limited by RAM bandwidth, not by the GPU.

Where to buy a RX 6800 XT

Prices and links follow the market you choose.

Radeon RX 6800 XT

16 GB GDDR6 · 512 GB/s · 300 W

RX 6800 XT 16 GB · 64 GB RAM → gpt-oss-20b

Price not available

Availability unknown

View on amazon.com

Affiliate link

No current offer imported for this market: the link opens a marketplace search.

Frequently asked questions

Which LLMs can the RX 6800 XT run?
Entirely on the GPU: Qwen3 30B-A3B (Q3_K_M), Gemma 3 27B IT (Q3_K_M), Mistral Small 3.2 24B Instruct (IQ4_XS), gpt-oss-20b (MXFP4), Qwen3 14B (Q6_K), DeepSeek R1 Distill Qwen 14B (Q6_K), Qwen2.5 14B Instruct (Q6_K), Phi-4 14B (Q6_K). Larger models need CPU/GPU offload.
Can the RX 6800 XT run a 70B model?
Not recommended: a 70B model at Q4_K_M needs about 45 GB; the RX 6800 XT has 16 GB.
How is the speed estimated?
Token generation reads every active weight once per token, so it is limited by memory bandwidth (512 GB/s here). We apply an efficiency range per backend and label the result as an estimate until real benchmarks exist.