AMD · unified x86

AMD Ryzen AI Max+ 395 (Radeon 8060S)

AMD Ryzen AI Max+ 395 (Radeon 8060S) has 32 GB of VRAM at 256 GB/s — about 22.32 GiB usable after driver and compositor overhead. 1881 of 2118 indexed models fit at 64K context with q8_0 KV. Note only 24 GB of its 32 GB is allocatable to the GPU.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed
Memory
32 GB
LPDDR5X-8000
Bandwidth
256 GB/s
256-bit bus
Tensor FP16
—
dense
TDP
120 W
Context4K8K16K32K64K128K
KV cachef16q8_0q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context
text 1604vision language 173image 2audio tts 21audio asr 39video 16embedding 26

What fits at 64K context

largest quantization that fits, per model · 1881 of 2118 indexed
ModelBest quantParamsWeights●KV●Total in memory◐Headroom◐tok/s◐
WhiteRabbitNeo-33B-v1Q3_K_M33.3B14.99 GiB8.23 GiB24.00 GiB0.00 GiB6±25%
Gemma-4-31B-Isometry-RPI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Gemma-4-Dark-Gemistry-31BI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Prosopon-31BI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Gemma-4-Novelist-Eclipse-31BI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Giftige-Blume-31B-v1-StyleSwapI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
G4-MeroMero-31B-StyleSwapI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Gemma-4-31B-StyleTune-heretic-araI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Pantheon-Reasoning-31B-1.1I1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Gemma-4-31B-StyleTuneI1-Q4_K_S32.7B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
Barcenas-StyleTune-31B-FableI1-Q4_K_S32.1B17.28 GiB5.94 GiB24.00 GiB0.00 GiB6±25%
MN-GRAND-23.5B-Gutenberg-UNCENSORED-V2-GLM4.7-ThinkingI1-Q4_023.4B12.49 GiB10.76 GiB23.99 GiB0.01 GiB6±25%
deepseek-coder-33b-baseQ3_K_M33.3B14.97 GiB8.23 GiB23.98 GiB0.02 GiB6±25%
GLM-Z1-Rumination-32B-0414Q3_K_M33.1B15.07 GiB8.10 GiB23.96 GiB0.04 GiB6±25%
Open_Gpt4_8x7B_v0.1MoEQ3_K_M46.7B18.97 GiB4.25 GiB23.95 GiB0.05 GiB8±37%
dolphin-2.7-mixtral-8x7bMoEQ3_K_M46.7B18.96 GiB4.25 GiB23.95 GiB0.05 GiB8±37%
Mixtral-8x7B-MoE-RP-StoryMoEQ3_K_M46.7B18.96 GiB4.25 GiB23.95 GiB0.05 GiB8±37%
v6-Finch-7B-HFQ6_K7.6B6.19 GiB17.00 GiB23.93 GiB0.07 GiB6±25%
rwkv-6-world-7bQ6_K7.6B6.19 GiB17.00 GiB23.93 GiB0.07 GiB6±25%
EXAONE-4.0-32BQ5_K_S32.0B20.56 GiB2.57 GiB23.93 GiB0.07 GiB6±25%
Smilodon-9B-v1F1610.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
bella-bartender-v2F169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
Gemma-The-Writer-9B-HERETIC-Uncensored-AbliteratedF169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
Gemma-2-9B-It-SPPO-Iter3BF169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
G2-Darkest-Writer-Dirty-Shirley-9B-v2F169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
G2-Darkest-Writer-9B-v1F169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
Gemma-SEA-LION-v3-9B-ITF169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
Tiger-Gemma-9B-v3F169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
gemma-2-9b-itF169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
magnum-v4-9bF169.2B17.22 GiB5.97 GiB23.93 GiB0.07 GiB6±25%
Fallen-Gemma3-27B-v1Q6_K_L27.4B20.96 GiB2.23 GiB23.93 GiB0.07 GiB6±25%
IQuest-Coder-V1-40B-InstructI1-IQ2_M39.8B12.50 GiB10.63 GiB23.92 GiB0.08 GiB6±25%
Skyfall-31B-v4.2IQ4_XS31.4B15.93 GiB7.17 GiB23.92 GiB0.08 GiB6±25%
Qwen3-Coder-NextMoEUD-IQ1_S79.7B20.03 GiB3.19 GiB23.91 GiB0.09 GiB17±37%
Nemotron-Labs-Audex-30B-A3BQ4_K_M32.0B23.16 GiB0.00 GiB23.90 GiB0.10 GiB6±25%
spoomplesmaxx-v2.1-30BIQ4_XS28.9B14.59 GiB8.50 GiB23.90 GiB0.10 GiB6±25%
granite-4.1-30b-hereticIQ4_XS28.9B14.59 GiB8.50 GiB23.90 GiB0.10 GiB6±25%
Skyfall-31B-v4.2-hereticIQ4_XS31.4B15.89 GiB7.17 GiB23.88 GiB0.12 GiB6±25%
gemma-4-31B-itQ4_031.3B17.16 GiB5.94 GiB23.88 GiB0.12 GiB6±25%
Darwin-35B-A3B-OpusMoEQ5_K_S36.0B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
Aurora-Code-1MoEQ5_K_S34.7B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
grug-35b-v2MoEQ5_K_S35.1B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
grug-35bMoEQ5_K_S35.1B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
WorldSim-Opus-3.6-35B-A3BMoEQ5_K_S35.1B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
Qwen3.6-35B-A3B-AnkoMoEQ5_K_S35.1B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
KAT-Coder-V2.5-DevMoEQ5_K_S34.7B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
Ornith-1.0-35BMoEQ5_K_S34.7B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
Nex-N2-miniMoEQ5_K_S35.1B22.50 GiB0.66 GiB23.87 GiB0.13 GiB29±37%
TildeOpen-30B-Instruct-LVI1-Q3_K_L30.7B15.11 GiB7.97 GiB23.86 GiB0.14 GiB6±25%
Qwen2.5-Coder-14B-InstructQ4_K_M14.8B16.74 GiB6.38 GiB23.86 GiB0.14 GiB6±25%
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16UD-Q5_K_S33.0B23.10 GiB0.00 GiB23.85 GiB0.15 GiB6±25%
OLMo-2-0325-32BQ3_K_M32.2B14.53 GiB8.50 GiB23.83 GiB0.17 GiB6±25%
granite-34b-code-base-8kI1-Q5_K_M33.7B23.05 GiB0.00 GiB23.83 GiB0.17 GiB6±25%
Rocinante-XL-16B-v1Q8_016.1B15.91 GiB7.17 GiB23.82 GiB0.18 GiB6±25%
Qwen3.5-40B-Claude-4.6-Opus-Deckard-Heretic-Uncensored-ThinkingIQ4_XS39.5B19.87 GiB3.19 GiB23.82 GiB0.18 GiB6±25%
Goetia-26B-A4B-v1.4MoEI1-Q6_K26.0B21.65 GiB1.48 GiB23.82 GiB0.18 GiB6±25%
G4-Moonlight-Dusk-26B-A4B-hereticMoEI1-Q6_K26.5B21.65 GiB1.48 GiB23.82 GiB0.18 GiB6±25%
Pantheon-Reasoning-26B-A4B-1.1-hereticMoEI1-Q6_K26.5B21.65 GiB1.48 GiB23.82 GiB0.18 GiB6±25%
G4-Moonlight-Dusk-26B-A4BMoEI1-Q6_K26.5B21.65 GiB1.48 GiB23.82 GiB0.18 GiB6±25%
Chimera-X-26B-A4BMoEI1-Q6_K26.5B21.65 GiB1.48 GiB23.82 GiB0.18 GiB6±25%
●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 AMD Ryzen AI Max+ 395 (Radeon 8060S) run?
1881 of 2118 indexed open-weight models fit a AMD Ryzen AI Max+ 395 (Radeon 8060S) at 65,536 context with q8_0 KV cache, the largest being WhiteRabbitNeo-33B-v1 at Q3_K_M. That covers text, vision-language, image, video and speech models.
How much usable memory does a AMD Ryzen AI Max+ 395 (Radeon 8060S) actually have?
Its nameplate is 32 GB, but about 22.32 GiB is available to a model once driver and compositor overhead is accounted for, and only 24 GB of the pool can be allocated to the GPU at all.
Is a AMD Ryzen AI Max+ 395 (Radeon 8060S) fast for local AI?
Its memory bandwidth is 256 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.