RX 6800
12 records across 5 models (quants: F16 (unquantized), GGUF F16, IQ4_XS, MXFP4, Q4_0, Q4_K_M).
| hardware | model | quant | backend | tok/s | pp tok/s | W | ttft s | ctx/tg | date | class | flags | source |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| RX 6800 | Gemma 4 12B | Q4_K_M | llama.cpp (Vulkan) | 52.1 | 799 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | Gemma 4 12B | Q4_K_M | llama.cpp (ROCm) | 48.9 | 736 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | gpt-oss 20B 23.6B | MXFP4 | llama.cpp (Vulkan) | 145.9 | 1883 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | gpt-oss 20B 23.6B | MXFP4 | llama.cpp (ROCm) | 121.2 | 2022 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 106.49 | 1597.95 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | llama 7B | Q4_0 | llama.cpp (Vulkan) | 101.47 | 1593.97 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | llama 7B | Q4_0 | llama.cpp (ROCm) | 93.6 | 1738.62 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | llama 7B | Q4_0 | llama.cpp (ROCm) | 86.03 | 1510.13 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | Phi-3 mini 4k instruct 3.8B | GGUF F16 | llama.cpp | 42 | – | – | – | – | 2024-07 | community | – | source |
| RX 6800 | Phi-3 mini 4k instruct 3.8B | F16 (unquantized) | ExLlamaV2 | 18 | – | – | – | – | 2024-07 | community | – | source |
| RX 6800 | Qwen3.8-27B | IQ4_XS | llama.cpp (Vulkan) | 28.4 | 356 | – | – | tg128 | 2026-10-01 | community | – | source |
| RX 6800 | Qwen3.8-27B | IQ4_XS | llama.cpp (ROCm) | 26.1 | 395 | – | – | tg128 | 2026-10-01 | community | – | source |
All rows cite their source. Provenance: community 40018 sourced 1085