Hardware

Radeon RX 6700 XT

Radeon RX 6700 XT has 12 GB GDDR6 and 384 GB/s. See which open models fit, at which quant, and roughly how fast. Figures below use the default memory (12 GB), 32 GB system RAM, and DDR5-5600 for anything that spills. Updated 2026-10-17.

Memory12 GB GDDR6
Bandwidth384 GB/s
Kindgpu
Year2021

12GB GDDR6, 192-bit, 384 GB/s. Source.

RSS for new models on this device

Models that fit

ModelVerdictQuantContexttok/s
Qwen3 14B
14.8B
Tight fitQ5_K_S4K19–26
NousResearch Hermes 4 14B
14.8B
Tight fitQ5_K_S4K19–26
Ministral 3 14B Instruct 2512
13.5B
Tight fitUD-Q5_K_XL8K19–26
Ministral 3 14B Reasoning 2512
13.5B
Tight fitUD-Q5_K_XL8K19–26
Qwen2.5 14B Instruct
14.8B
Tight fitQ4_K_M8K20–27
DeepSeek R1 Distill Qwen 14B
14.8B
Tight fitQ4_K_M8K20–27
Qwen2.5 Coder 14B Instruct 128K
14.8B
Tight fitQ4_K_M8K20–27
phi 4
14.7B
Tight fitQ4_K_M8K20–27
gemma 4 12b it
11.9B
Tight fitQ6_K8K16–22
Devstral Small 2 24B Instruct 2512
23.6B
Tight fitUD-IQ3_XXS8K20–27
Mistral Small 3.2 24B Instruct 2506
23.6B
Tight fitUD-IQ3_XXS8K20–27
Magistral Small 2509
23.6B
Tight fitUD-IQ3_XXS8K20–27
Qwen3.8 27B
27.3B
Tight fitUD-Q2_K_XL8K20–28
Qwen2.5 32B Instruct
32.8B
Tight fitIQ2_XXS4K21–29
Qwen2.5 Coder 32B Instruct
32.8B
Tight fitIQ2_XXS4K21–29
Qwen3 32B
32.8B
Tight fitUD-IQ2_XXS4K20–28
Qwen3 VL 32B Instruct
32.8B
Tight fitUD-IQ2_XXS4K20–28
GLM 4 32B 0414
32.6B
Tight fitUD-IQ2_XXS8K22–29
GLM 4.6V Flash
9.4B
Tight fitQ8_08K20–28
GLM Z1 9B 0414
9.4B
Tight fitQ8_08K20–28
GLM 4 9B 0414
9.4B
Tight fitQ8_08K20–28
GLM 4.1V 9B Thinking
9.4B
Tight fitQ8_08K20–28
gemma 4 31B it
30.7B
Tight fitUD-IQ2_XXS8K13–17
Qwen3.5 9B
9.0B
Tight fitQ8_08K22–29
MiMo V2.6 Distill Qwen 9B
9.0B
Tight fitQ8_08K22–29
Ministral 3 8B Instruct 2512
8.5B
Tight fitQ8_08K21–28
Ministral 3 8B Reasoning 2512
8.5B
Tight fitQ8_08K21–28
Qwen3.5 27B
26.9B
RunsUD-IQ2_XXS8K23–32
Qwen3.6 27B
26.9B
Tight fitUD-IQ2_XXS8K21–29
Qwen3 VL 8B Instruct
8.2B
Tight fitQ8_08K21–29
DeepSeek R1 0528 Qwen3 8B
8.2B
Tight fitQ8_08K21–29
Qwen3 8B
8.2B
Tight fitQ8_08K21–29
Qwen3 VL 8B Thinking
8.2B
Tight fitQ8_08K21–29
gemma 3 27b it
27.0B
RunsUD-IQ2_XXS8K18–24
medgemma 27b text it
27.0B
RunsUD-IQ2_XXS8K18–24
Llama 3.1 8B Instruct
8.0B
RunsUD-Q6_K_XL8K25–34
DeepSeek R1 Distill Llama 8B
8.0B
Tight fitQ8_08K22–30
Apertus 8B Instruct 2509
8.1B
Tight fitQ8_08K22–30
GLM 4.7 Flash REAP 23B A3B
23.0B
Tight fitQ3_K_S8K118–161
Kwaipilot KAT Coder V2.5 Dev
34.7B
Tight fitIQ2_XXS8K109–148

90 more models in the catalog do not fit this default profile, or ranked below the first 40 that do.

Example: Tight fit fully on your Radeon RX 6700 XT at Q5_K_S, 4K context, about 19–26 tok/s. Weights 9.56 GB plus 0.63 GB of KV cache — 10.6 GB of 10.9 GB usable. Weights shown in the verdict use GiB (the same 1024³ counting as GPU memory).