What AI models can a Arc Pro B60 24GB run?

1872 of 2118 indexed open-weight models fit a Arc Pro B60 24GB at 65,536 context with q8_0 KV cache, the largest being solar-pro-preview-instruct at Q3_K_L. That covers text, vision-language, image, video and speech models.

How much usable memory does a Arc Pro B60 24GB actually have?

Its nameplate is 24 GB, but about 22.32 GiB is available to a model once driver and compositor overhead is accounted for.

Is a Arc Pro B60 24GB fast for local AI?

Its memory bandwidth is 456 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.

Intel · workstation

Arc Pro B60 24GB

Arc Pro B60 24GB has 24 GB of VRAM at 456 GB/s — about 22.32 GiB usable after driver and compositor overhead. 1872 of 2118 indexed models fit at 64K context with q8_0 KV.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed

Memory

24 GB

GDDR6

Bandwidth

456 GB/s

192-bit bus

Tensor FP16

99 TF

dense

TDP

200 W

$599 MSRP

Context4K 8K 16K 32K 64K 128K

KV cachef16 q8_0 q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context

text 1597vision language 171audio asr 39image 2audio tts 21video 16embedding 26

What fits at 64K context

largest quantization that fits, per model · 1872 of 2118 indexed

Model	Best quant	Params	Weights●	KV●	Total in memory◐	Headroom◐	tok/s◐
solar-pro-preview-instructKV unresolved	Q3_K_L	22.1B	10.84 GiB	10.63 GiB	22.32 GiB	0.00 GiB	12±30%
ThinkingCap-Qwen3.6-27B	Q5_K_M	27.4B	19.33 GiB	2.13 GiB	22.32 GiB	0.00 GiB	12±30%
Tess-4-27B	Q5_K_M	27.8B	19.33 GiB	2.13 GiB	22.32 GiB	0.00 GiB	12±30%
Snowpiercer-15B-v4-heretic	Q8_0	15.0B	14.83 GiB	6.64 GiB	22.31 GiB	0.01 GiB	12±30%
Snowpiercer-15B-v4	Q8_0	15.0B	14.83 GiB	6.64 GiB	22.31 GiB	0.01 GiB	12±30%
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16	UD-Q4_K_S	33.0B	21.47 GiB	0.00 GiB	22.31 GiB	0.01 GiB	12±30%
Qwen3.6-35B-A3B-Fable-5-DistillMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Qwable-v2MoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Salience-1.5-ProMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Qwen3.6-35B-A3B-YOYO-V2MoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Ornith-1.0-35B-FP8-BLOCK-MTPMoE	I1-Q4_1	35.5B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
fable-coder-35B-A3BMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Qwen3.6-35B-A3B-AntiLoopMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
PINQWEN-3.6-35B-CLEAN-BF16MoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
UniMath-35B-A3BMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Ornith-1.0-35B-Heretic-MTPMoE	I1-Q4_1	—	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Fawen-1.0-35BMoE	I1-Q4_1	36.0B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
Qwen3.5-35B-A3B-uncensored-heretic-v2-Native-MTP-PreservedMoE	I1-Q4_1	35.1B	20.84 GiB	0.66 GiB	22.31 GiB	0.01 GiB	52±37%
SambaLingo-Japanese-Chat	I1-Q5_K_S	6.9B	4.48 GiB	17.00 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Gembrain-X-Core-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Gembrain-X-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-31B-Isometry-Fabled-Persona	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Versipellis-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma4-Gutenberg-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
G4-MeroMero-31B-uncensored-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Novelist-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Wanabi-Gemma4-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
G4-Alice-v1.2-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Agares-31B-v1	I1-Q3_K_L	30.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma4-Gutenberg-31B-Heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-Ortenzya-The-Creative-Wordsmith-31B-it-uncensored-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Gemsicle-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Gembrain-31B-it-uncensored-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Melinoe-Gemma4-31B-VL-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
G4-MeroMero-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Glistening-Gem-31B-v1.0	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Melinoe-Gemma4-31B-VL	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-31B-Storymaxxed3	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Huihui-gemma-4-31B-it-qat-q4_0-unquantized-abliterated	I1-Q3_K_L	32.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-Queen-it-qat-q4_0-unquantized	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-it-qat-q4_0-unquantized-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-AssGuard-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
copywriter-gemma4-31b	I1-Q3_K_L	32.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-heretic-finetune	I1-Q3_K_L	30.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Garnet-V2-31B-it-ultra-uncensored-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-it-abliterated-v3	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Harmonia-31B-uncensored-heretic	Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-it-noloop	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Webs-Sejong-31B-v7	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Lilith-31B-v1.0	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
JGOS-31B-Think	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-Mergemaxxed	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
K1-v6-zero	I1-Q3_K_L	32.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-it-uncensored-heretic	Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Queen-31B-it-uncensored-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-Sphinsikus-Chronist-31B	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
gemma-4-31B-it-heretic	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma4-31B-Finetuned-V2	I1-Q3_K_L	32.7B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-31B-storymaxxed	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±30%
Gemma-4-31B-storymaxxed2	I1-Q3_K_L	31.3B	15.49 GiB	5.94 GiB	22.30 GiB	0.02 GiB	12±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 Pro B60 24GB run?: 1872 of 2118 indexed open-weight models fit a Arc Pro B60 24GB at 65,536 context with q8_0 KV cache, the largest being solar-pro-preview-instruct at Q3_K_L. That covers text, vision-language, image, video and speech models.
How much usable memory does a Arc Pro B60 24GB actually have?: Its nameplate is 24 GB, but about 22.32 GiB is available to a model once driver and compositor overhead is accounted for.
Is a Arc Pro B60 24GB fast for local AI?: Its memory bandwidth is 456 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.