Intel · consumer

Arc B570 10GB

Arc B570 10GB has 10 GB of VRAM at 380 GB/s — about 9.30 GiB usable after driver and compositor overhead. 1429 of 2118 indexed models fit at 32K context with q8_0 KV.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed
Memory
10 GB
GDDR6
Bandwidth
380 GB/s
160-bit bus
Tensor FP16
—
dense
TDP
150 W
$219 MSRP
Context4K8K16K32K64K128K
KV cachef16q8_0q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context
text 1220vision language 111image 1video 12embedding 26audio asr 38audio tts 21

What fits at 32K context

largest quantization that fits, per model · 1429 of 2118 indexed
ModelBest quantParamsWeights●KV●Total in memory◐Headroom◐tok/s◐
OmniAtlas-Qwen3-30B-A3BI1-IQ2_XS31.7B8.45 GiB0.00 GiB9.30 GiB0.00 GiB24±30%
Qwen3-Omni-30B-A3B-CaptionerI1-IQ2_XS31.7B8.45 GiB0.00 GiB9.30 GiB0.00 GiB24±30%
DeepSeek-Coder-V2-Lite-BaseMoEI1-IQ4_XS15.7B7.98 GiB0.50 GiB9.30 GiB0.00 GiB63±37%
DeepSeek-V2-Lite-ChatMoEIQ4_XS15.7B7.98 GiB0.50 GiB9.30 GiB0.00 GiB63±37%
Qwen3-Coder-REAP-25B-A3BMoEIQ2_XS24.9B6.91 GiB1.59 GiB9.30 GiB0.00 GiB44±37%
Grug-12BQ4_K_M12.0B7.14 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-Esper4Q4_K_M12.0B7.14 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-itQ4_K_M12.0B7.14 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Qwen3.6-28BMoEI1-IQ2_S28.2B8.16 GiB0.33 GiB9.29 GiB0.01 GiB89±37%
Qwen3.5-28BMoEI1-IQ2_S28.7B8.16 GiB0.33 GiB9.29 GiB0.01 GiB89±37%
NVIDIA-Nemotron-Nano-9B-v2IQ3_XXS8.9B4.73 GiB3.72 GiB9.29 GiB0.01 GiB24±30%
HomunculusQ3_K_M12.5B5.79 GiB2.66 GiB9.29 GiB0.01 GiB24±30%
SuperGemma-4-12b-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-agentic-fable5-composer2.5-v2-3.5x-tau2-uncensored-hereticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-coder-fable5-composer2.5-v1-uncensored-hereticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-uncensored-hereticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Aura-Medium-v1-BF16I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-GuardpointI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Gemma-4-12B-it-AEON-Abliterated-K4-BF16I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-Tachibana-AgentI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-marvin-gutenberg-rp-v2I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-crownelius-writerI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Huihui-gemma-4-12B-coder-fable5-composer2.5-v1-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-asterion-agenticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Huihui-gemma-4-12B-agentic-fable5-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
g4-12b-it-trismegistusI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma4-12b-it-asimovI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
FabGemmaI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Huihui-gemma-4-12B-it-qat-q4_0-unquantized-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-abliterated-uncensoredI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Gemma-4-12b-it-AbliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-Queen-it-qat-q4_0-unquantizedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-heretic_decensoredI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Iris-12B-gemma-4-it-qatI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-coder-fable5-composer2.5-v1I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
G4-Starry-Ocean-12BI1-Q4_111.9B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-QAT-SOMPOA-heresyI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-uncensored-opus4.7-cotI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Gemma4-12B-IT-AbliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-it-uncensoredI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Huihui-gemma-4-12B-it-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-it-hereticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Tema_Q-X5-12B-ThinkingI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12B-coder-fable5-composer2.5-v1-bf16I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
swarm-sovereign-12bI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Gemma-4-12B-OBLITERATEDI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Gemma4-12B-UncensoredI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Serenity-12BI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Dark-PaneI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Reelva-12BI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
G4-Starry-Ocean-12B-hereticI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Iris-12B-v1.3.2I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Semancer-12BI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
Iris-12B-v1.2I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-heretic-abliteratedI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-marvin-gutenbergI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-marvin-v2I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
gemma-4-12b-marvin-v1I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
STARK-WEB-12B-v1.7I1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
STARK-WEB-12BI1-Q4_112.0B7.13 GiB1.31 GiB9.29 GiB0.01 GiB24±30%
●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 Arc B570 10GB run?
1429 of 2118 indexed open-weight models fit a Arc B570 10GB at 32,768 context with q8_0 KV cache, the largest being OmniAtlas-Qwen3-30B-A3B at I1-IQ2_XS. That covers text, vision-language, image, video and speech models.
How much usable memory does a Arc B570 10GB actually have?
Its nameplate is 10 GB, but about 9.30 GiB is available to a model once driver and compositor overhead is accounted for.
Is a Arc B570 10GB fast for local AI?
Its memory bandwidth is 380 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.