AMD · consumer

Radeon RX 6400

Radeon RX 6400 has 4 GB of VRAM at 128 GB/s — about 3.72 GiB usable after driver and compositor overhead. 823 of 2118 indexed models fit at 4K context with q8_0 KV.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed
Memory
4 GB
GDDR6
Bandwidth
128 GB/s
64-bit bus
Tensor FP16
—
dense
TDP
53 W
$159 MSRP
Context4K8K16K32K64K128K
KV cachef16q8_0q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context
audio tts 19vision language 58text 681embedding 25image 1audio asr 37video 2

What fits at 4K context

largest quantization that fits, per model · 823 of 2118 indexed
ModelBest quantParamsWeights●KV●Total in memory◐Headroom◐tok/s◐
orpheus-3b-0.1-pretrainedQ5_K_S3.8B2.58 GiB0.23 GiB3.72 GiB0.00 GiB28±26.5%
gemma-3n-E2B-itQ4_K_S5.4B2.77 GiB0.05 GiB3.72 GiB0.00 GiB28±26.5%
Vikhr-Gemma-2B-instructQ8_02.6B2.59 GiB0.22 GiB3.72 GiB0.00 GiB28±26.5%
gemma-2-2b-it-abliteratedQ8_02.6B2.59 GiB0.22 GiB3.72 GiB0.00 GiB28±26.5%
gemma-2-2b-itQ8_02.6B2.59 GiB0.22 GiB3.72 GiB0.00 GiB28±26.5%
Gemmasutra-Mini-2B-v1Q8_02.6B2.59 GiB0.22 GiB3.72 GiB0.00 GiB28±26.5%
OvisOCR2F32853M2.81 GiB0.02 GiB3.72 GiB0.00 GiB28±26.5%
Phi-4-mini-reasoningQ5_K_S3.8B2.54 GiB0.27 GiB3.72 GiB0.00 GiB28±26.5%
Phi-4-mini-instructQ5_K_S3.8B2.54 GiB0.27 GiB3.72 GiB0.00 GiB28±26.5%
deepseek-coder-6.7B-kexerI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.72 GiB0.00 GiB29±26.5%
Magicoder-S-DS-6.7BI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.72 GiB0.00 GiB29±26.5%
deepseek-coder-6.7b-baseI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.72 GiB0.00 GiB29±26.5%
WizardLM-7B-UncensoredI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.71 GiB0.01 GiB29±26.5%
Llama-2-7B-32K-InstructI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.71 GiB0.01 GiB29±26.5%
Luna-AI-Llama2-UncensoredI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.71 GiB0.01 GiB29±26.5%
Swallow-7b-NVE-instruct-hfI1-IQ2_XXS6.7B1.73 GiB1.06 GiB3.71 GiB0.01 GiB29±26.5%
nomic-embed-codeQ2_K7.1B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Nemotron-3-Embed-8B-BF16IQ2_S8.0B2.49 GiB0.28 GiB3.71 GiB0.01 GiB29±26.5%
DeepHat-V1-7B-Heretic-AbliteratedI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
ShizhenGPT-7B-VLI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
HuatuoGPT-o1-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
MathSmith-DS-Qwen-7B-LongCoTI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
AstraGPTCoder-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-Coder-7B-Instruct-Ghidra-v2I1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
EsDrac-v1-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Hemlock-Apothecary-7B-GRPO-e3I1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
openhands-lm-7b-v0.1I1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Hemlock2-Coder-7B-GRPOI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
shellwhiz-7bI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-Coder-7B-Instruct-OBLITERATED-advancedI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen-STEM-Specialist-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
VulnLLM-R-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Garnet-OCR-7B-0422I1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
UwU-7B-InstructI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Video-R1-7BI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
HARC-Qwen2.5-7B-InstructI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-Coder-7B-AbliteratedI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Bozdogan-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Crazy-AI-ModelI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
turbo-ai-7bI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
DeepSeek-R1-Distill-Qwen-7B-abliterated-v2I1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Ghosty-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
SP-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-Coder-7B-Instruct-UncensoredI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-VL-7B-Instruct-abliteratedI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
DeepSeek-R1-Distill-Qwen-8B-AbliteratedI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-7BQ2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Qwen2.5-VL-7B-Instruct-hereticI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
AWARES-Qwen2.5-VL-7BI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
olmOCR-2-7B-1025I1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Med-RwRI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
SpatialThinker-7BI1-Q2_K_S8.3B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
MQ-Coldbrew-BaseI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Kepler-Reasoning-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
DeepSeek-R1-Distill-Qwen-7B-Uncensored-ReasonerI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
DeepSeek-R1-Distill-Qwen-7B-UncensoredI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
DeepSeek-R1-STEM-Coder-7BI1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Kepler-8B-Instruct-v2I1-Q2_K_S7.6B2.64 GiB0.12 GiB3.71 GiB0.01 GiB29±26.5%
Maestro1-9BTQ2_08.8B2.48 GiB0.30 GiB3.71 GiB0.01 GiB29±26.5%
gemma-4-E2B-it-ultra-uncensored-hereticI1-Q2_K5.1B2.78 GiB0.03 GiB3.71 GiB0.01 GiB28±26.5%
●From the file◐Predictedwhat these mean

Speed is modeled, not measured: decode is memory-bandwidth bound, so tokens per second is bytes read per token against achievable bandwidth. Mixture-of-experts models carry a wider band because only the routed experts are read each step, and few have been measured publicly.

Questions people ask

What AI models can a Radeon RX 6400 run?
823 of 2118 indexed open-weight models fit a Radeon RX 6400 at 4,096 context with q8_0 KV cache, the largest being orpheus-3b-0.1-pretrained at Q5_K_S. That covers text, vision-language, image, video and speech models.
How much usable memory does a Radeon RX 6400 actually have?
Its nameplate is 4 GB, but about 3.72 GiB is available to a model once driver and compositor overhead is accounted for.
Is a Radeon RX 6400 fast for local AI?
Its memory bandwidth is 128 GB/s, and that figure — not teraflops — is what governs token generation speed. Capacity decides what you can run; bandwidth decides how fast it runs.
Radeon RX 6400 — what AI models can it run locally? — ossmodeldb