RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

What AI models run on the Arc A770 16GB?

The Arc A770 16GB has 16 GB of VRAM (560 GB/s). About 15.2 GB is usable for a model. 194 of 285 open models we track fit at 8K context.

194models fit (8K ctx)
180still fit at 32K ctx
40Blargest that fits
560 GB/smemory bandwidth

This card has 16 GB of VRAM. See every model that fits in 16 GB VRAM. Other 16 GB hardware: RTX 4060 Ti 16GB, RTX 4070 Ti Super, RTX 4080, RTX 4080 Super, RTX 5060 Ti 16GB, RTX 5070 Ti, RTX 5080, RX 7800 XT.

Models that fit, largest first

ModelParamsBest quant (8K)Needs32K contextMax tok/s (est.)
flint40B (1.3B act.)Q2_K14.8 GBno1195details
Maxwell V1 Flash37B (2.3B act.)Q2_K14.0 GBQ2_K658details
Qwen3.6 35B A3B36B (3.0B act.)Q2_K13.6 GBQ2_K498details
Ornith 1.5 35B A3B36B (3.0B act.)Q2_K13.2 GBQ2_K513details
Qwen3.5 35B A3B36B (3.0B act.)Q2_K13.6 GBQ2_K498details
Axion 1.0 35B35B (3.9B act.)Q2_K13.3 GBQ2_K383details
Qwen AgentWorld 35B A3B35B (3.0B act.)Q2_K13.1 GBQ2_K498details
Yi 1.5 34B Chat34BQ2_K14.3 GBno44details
Nous Hermes 2 Yi 34B34BQ2_K14.8 GBno42details
DeepSeek R1 Distill Qwen 32B33BQ2_K14.2 GBno45details
Qwen3 32B33BQ2_K14.1 GBno46details
GLM Z1 32B 041433BQ2_K12.6 GBQ2_K46details
GLM 4 32B 041433BQ2_K12.6 GBQ2_K46details
granite 4.0 h small32B (8.8B act.)Q2_K11.7 GBQ2_K174details
gemma 4 31B it31BQ2_K13.9 GBno48details
GLM 4.7 Flash31B (4.6B act.)Q2_K11.7 GBQ2_K334details
Qwen3 Coder 30B A3B Instruct31B (3.0B act.)Q2_K11.9 GBQ2_K506details
North Mini Code 1.030B (3.3B act.)Q2_K12.0 GBQ2_K452details
Hy Memory 30B A3B30B (3.0B act.)Q2_K11.9 GBQ2_K498details
Muse Glimmer 30B30BQ2_K11.4 GBQ2_K50details
granite 4.2 30b29BQ2_K12.7 GBno52details
Qwen3.8 27B28BQ2_K11.1 GBQ2_K54details
Qwen3.6 27B28BQ2_K11.1 GBQ2_K54details
Qwen3.5 27B28BQ2_K11.1 GBQ2_K54details
gemma 3 27b it27BQ2_K11.7 GBQ2_K53details
Arete 27B27BQ3_K_M13.6 GBQ2_K43details
MiniMax H3 Image PE27BQ2_K11.3 GBQ2_K52details
Trinity Mini26B (2.0B act.)Q3_K_M12.3 GBQ3_K_M615details
GLiDE 27B26BQ3_K_M13.2 GBQ2_K44details
diffusiongemma 26B A4B it26B (4.0B act.)Q3_K_M13.2 GBQ2_K286details
gemma 4 26B A4B it26B (4.0B act.)Q3_K_M13.3 GBQ2_K284details
LFM2.5 24B A2B24B (2.0B act.)Q3_K_M11.7 GBQ3_K_M573details
ERNIE 4.5 21B A3B PT22BQ4_K_M13.5 GBQ3_K_M42details
ERNIE 4.5 21B A3B Thinking22BQ4_K_M13.5 GBQ3_K_M42details
gpt oss safeguard 20b22B (4.8B act.)Q8_012.3 GBQ8_0208details
gpt oss 20b21B (4.2B act.)Q8_012.3 GBQ8_0231details
Groot 19B A3B 26102319B (3.0B act.)Q4_K_M12.7 GBQ4_K_M291details
Kimi VL A3B Thinking 250616B (3.0B act.)Q6_K13.5 GBQ5_K_M228details
Kimi VL A3B Instruct16B (3.0B act.)Q5_K_M12.2 GBQ5_K_M254details
Kimi VL A3B Thinking16B (3.0B act.)Q6_K13.5 GBQ5_K_M228details
deepseek moe 16b base16B (2.8B act.)Q4_K_M12.5 GBQ2_K299details
LLaDA2.1 mini16B (1.4B act.)Q6_K13.4 GBQ5_K_M477details
Moonlight 16B A3B Instruct16B (3.0B act.)Q5_K_M12.2 GBQ5_K_M247details
DeepSeek Coder V2 Lite Instruct16B (2.7B act.)Q5_K_M11.9 GBQ5_K_M279details
DeepSeek V2 Lite16B (2.7B act.)Q5_K_M11.3 GBQ5_K_M295details
DeepSeek V2 Lite Chat16B (2.7B act.)Q5_K_M11.9 GBQ5_K_M279details
Qwen2.5 Coder 14B Instruct15BQ6_K13.5 GBQ3_K_M46details
DeepSeek R1 Distill Qwen 14B15BQ6_K13.5 GBQ3_K_M46details
Qwen3 14B15BQ6_K13.2 GBQ3_K_M46details
phi 415BQ6_K13.4 GBQ3_K_M47details
gemma 3 12b it12BQ8_013.4 GBQ6_K45details
Llama Guard 4 12B12BQ6_K11.5 GBQ4_K_M57details
Llama 3.2 11B Vision Instruct11BQ8_011.6 GBQ5_K_M54details
Llama 3.2 11B Vision11BQ8_012.4 GBQ5_K_M50details
GLM 4.6V Flash10BQ8_010.3 GBQ8_056details
GLM 4.1V 9B Thinking10BQ8_010.3 GBQ8_056details
Step3 VL 10B10BQ8_09.9 GBQ8_064details
SenseNova Loop 9B Preview9.7BQ8_010.0 GBQ8_059details
Qwen3.5 9B9.7BQ8_010.0 GBQ8_059details
Ornith 1.5 9B9.7BQ8_010.0 GBQ8_059details
Qwen3.5 9B Base9.7BQ8_010.0 GBQ8_059details
MiMo V2.6 Distill Qwen 9B9.4BQ8_010.0 GBQ8_059details
glm 4 9b9.4BQ6_K13.4 GBno68details
internlm3 8b instruct8.8BQ8_09.7 GBQ8_060details
granite 4.2 8b8.8BQ8_010.5 GBQ6_K60details
granite 4.1 8b8.8BQ8_010.5 GBQ6_K60details
Qwen3 VL 8B Instruct8.8BQ8_09.9 GBQ8_064details
Molmo2 8B8.7BQ8_09.9 GBQ8_064details
LFM2.5 8B A1B8.5B (1.0B act.)Q8_09.1 GBQ8_0526details
Qwen2.5 VL 7B Instruct8.3BQ8_08.7 GBQ8_069details
UI TARS 1.5 7B8.3BQ8_08.7 GBQ8_069details
olmOCR 2 7B 10258.3BQ8_08.7 GBQ8_069details
olmOCR 7B 0225 preview8.3BQ8_08.7 GBQ8_069details
UI TARS 7B DPO8.3BQ8_08.7 GBQ8_069details
Qwen3 8B8.2BQ8_09.9 GBQ8_064details
DeepSeek R1 0528 Qwen3 8B8.2BQ8_09.9 GBQ8_064details
granite 3.3 8b instruct8.2BQ8_09.8 GBQ8_064details
granite 3.1 8b instruct8.2BQ8_09.8 GBQ8_064details
granite 3.2 8b instruct8.2BQ8_09.8 GBQ8_064details
Apertus 8B Instruct 25098.1BQ8_09.6 GBQ8_065details
Llama 3.1 8B Instruct8.0BQ8_09.6 GBQ8_066details
Llama 3.1 8B8.0BQ8_09.6 GBQ8_066details
Hermes 3 Llama 3.1 8B8.0BQ8_09.6 GBQ8_066details
DeepSeek R1 Distill Llama 8B8.0BQ8_09.6 GBQ8_066details
Meta Llama 3 8B Instruct8.0BQ8_09.6 GBQ8_066details
Meta Llama Guard 2 8B8.0BQ8_09.6 GBQ8_066details
Meta Llama 3.1 8B Instruct8.0BQ8_09.6 GBQ8_066details
Ling 3.0 tiny7.9B (1.4B act.)Q8_08.7 GBQ8_0380details
MiMo 7B RL7.8BQ8_09.4 GBQ8_069details
internlm2 5 7b chat7.7BQ8_09.2 GBQ8_068details
internlm2 chat 7b7.7BQ8_09.2 GBQ8_068details
Qwen2.5 7B Instruct7.6BQ8_08.7 GBQ8_069details
Qwen2.5 Coder 7B Instruct7.6BQ8_08.7 GBQ8_069details
DeepSeek R1 Distill Qwen 7B7.6BQ8_08.7 GBQ8_069details
Falcon H1R 7B7.6BQ8_08.5 GBQ8_069details
Olmo 3 7B Instruct7.3BQ8_010.3 GBQ8_072details
Olmo 3 1025 7B7.3BQ8_010.3 GBQ8_072details
Olmo 3 7B Think7.3BQ8_010.3 GBQ8_072details
Mistral 7B v0.17.2BQ8_08.6 GBQ8_073details
Mistral 7B Instruct v0.27.2BQ8_08.6 GBQ8_073details
Mistral 7B Instruct v0.17.2BQ8_08.6 GBQ8_073details
falcon 7b7.2BQ8_012.1 GBno73details
falcon 7b instruct7.2BQ8_011.8 GBno76details
granite 4.0 h tiny6.9B (1.5B act.)F1613.6 GBQ8_0191details
OLMoE 1B 7B 09246.9B (1.3B act.)Q8_08.4 GBQ8_0411details
deepseek coder 7b instruct v1.56.9BQ8_011.2 GBno76details
deepseek coder 6.7b instruct6.7BQ8_011.1 GBno78details
deepseek coder 6.7b base6.7BQ8_011.1 GBno78details
Llama 2 7b hf6.7BQ8_011.1 GBno78details
Yi 6B 200K6.1BF1612.3 GBQ8_046details
Yi 6B6.1BF1612.3 GBQ8_046details
Molmo2 4B4.9BF1610.0 GBF1663details
WeVisOne VL 4B4.8BF1610.0 GBF1663details
Mage VL Cowork4.7BF1610.0 GBF1663details
Qianfan OCR4.7BF1610.1 GBF1663details
Qwen3.5 4B4.7BF169.0 GBF1666details
S5 nano4.7BF169.2 GBF1665details
LightOnOCR 3 4B4.5BF169.0 GBF1666details
SingVL S1 4B4.5BF169.0 GBF1666details
Intern Decision 4B4.5BF169.0 GBF1666details
Qwen3 VL 4B Instruct4.4BF169.3 GBF1670details
gemma 3 4b it4.3BF168.4 GBF1672details
medgemma 1.5 4b it4.3BF168.4 GBF1672details
medgemma 4b it4.3BF168.4 GBF1672details
gemma 3 4b pt4.3BF169.2 GBF1665details
Phi 3.5 vision instruct4.1BF1611.2 GBQ4_K_M68details
Qwen3 4B4.0BF169.3 GBF1670details
Qwen3 4B Instruct 25074.0BF169.3 GBF1670details
granite vision 4.1 4b4.0BF167.6 GBF1682details
Phi 4 mini instruct3.8BF168.9 GBF1673details
Phi 3 mini 4k instruct3.8BF1610.6 GBQ5_K_M73details
Phi 3.5 mini instruct3.8BF1610.6 GBQ5_K_M73details
Phi 3 mini 128k instruct3.8BF1610.6 GBQ5_K_M73details
Qwen2.5 VL 3B Instruct3.8BF166.7 GBF1691details
granite 4.2 3b3.7BF168.0 GBF1676details
granite 4.1 3b llmon preview3.7BF168.0 GBF1677details
Raon 1 nano3.5BF167.0 GBF1689details
granite 4.1 3b3.4BF167.6 GBF1682details
granite 4.0 micro3.4BF167.6 GBF1682details
DeepSeek OCR 23.4B (1.2B act.)F166.6 GBF16270details
deepseek vl2 tiny3.4B (1.2B act.)F167.4 GBF16238details
Unlimited OCR 23.3B (1.1B act.)F166.6 GBF16277details
Unlimited OCR3.3B (1.1B act.)F166.6 GBF16279details
DeepSeek OCR3.3B (1.1B act.)F166.6 GBF16279details
granite 3.1 3b a800m instruct3.3B (883M act.)F167.1 GBF16317details
Llama 3.2 3B Instruct3.2BF1613.5 GBQ8_044details
Llama 3.2 3B3.2BF167.5 GBF1687details
Hermes 3 Llama 3.2 3B3.2BF167.5 GBF1687details
d1 3B3.1BF1610.8 GBF1652details
LFM2.5 VL 3B3.1BF1610.8 GBF1652details
Qwen2.5 3B Instruct3.1BF166.7 GBF1691details
SmolLM3 3B3.1BF166.9 GBF1691details
SmolLM3 3B Base3.1BF166.9 GBF1691details
phi 22.8BF168.2 GBQ8_0101details
LFM2.5 2.6B2.7BF1610.8 GBF1652details
granite 3.3 2b instruct2.5BF165.8 GBF16110details
MiniCPM5 2B2.5BF165.7 GBF16111details
WeVisOne VL 2B2.4BF165.4 GBF16137details
Qwen3.5 2B2.3BF164.5 GBF16148details
SmolVLM Instruct2.2BF165.4 GBF16154details
Qwen3 1.7B2.0BF164.8 GBF16162details
Qwen3.8 27B DSpine b162.0BF164.8 GBF16138details
d1 2B1.9BF167.2 GBF1681details
LFM3 VL 2B1.9BF167.2 GBF1681details
DeepSeek R1 Distill Qwen 1.5B1.8BF164.2 GBF16157details
LFM3 1.7B1.7BF167.2 GBF1681details
LFM3 1.7B Base1.7BF164.0 GBF16162details
SmolLM2 1.7B Instruct1.7BF165.2 GBF16164details
SmolLM2 1.7B1.7BF165.2 GBF16164details
SmolLM 1.7B1.7BF165.2 GBF16164details
granite 4.0 1b1.6BF164.3 GBF16171details
Qwen2.5 1.5B Instruct1.5BF163.8 GBF16181details
OLMo 2 0425 1B1.5BF164.4 GBF16188details
deepseek coder 1.3b instruct1.3BF164.5 GBF16208details
Llama 3.2 1B1.2BF163.2 GBF16227details
Llama 3.2 1B Instruct1.2BF165.5 GBF16113details
LFM2.5 1.2B Instruct1.2BF165.0 GBF16119details
HunyuanOCR1.1BF163.5 GBF16250details
PaddleOCR VL 1.6959MF162.5 GBF16292details
Qwen3.5 0.8B873MF162.4 GBF16369details
LightOnOCR 3 0.8B853MF162.4 GBF16369details
FastVLM 0.5B759MF162.2 GBF16369details
Qwen3 0.6B752MF162.7 GBF16467details
GOT OCR2 0716MF162.8 GBF16391details
Nemotron 3.5 Parse631MF161.9 GBF16428details
GOT OCR 2.0 hf561MF162.5 GBF16500details
SmolVLM2 500M Video Instruct507MF161.6 GBF16683details
SmolVLM 500M Instruct507MF161.6 GBF16683details
Qwen2.5 0.5B Instruct494MF161.7 GBF16567details
LFM2.5 VL 450M449MF161.9 GBF16394details
SmolLM2 360M Instruct362MF161.5 GBF16772details
SmolLM2 360M362MF161.5 GBF16772details
LFM2.5 350M354MF161.9 GBF16394details
LFM2 350M354MF161.4 GBF16790details

Just too big at 4-bit

These need up to 35% more memory than the Arc A770 16GB has at Q4_K_M and 8K context. A smaller quant or shorter context may work.

ModelQ4_K_M needsShort by
Trinity Mini15.8 GB0.6 GB
GLiDE 27B16.2 GB1.0 GB
MiniMax H3 Image PE16.8 GB1.6 GB
Qwen3.8 27B16.8 GB1.6 GB
Qwen3.5 27B17.0 GB1.8 GB
Qwen3.6 27B17.0 GB1.8 GB
diffusiongemma 26B A4B it17.1 GB1.9 GB
gemma 4 26B A4B it17.2 GB2.0 GB
gemma 3 27b it17.3 GB2.1 GB
Arete 27B17.4 GB2.2 GB
Muse Glimmer 30B17.9 GB2.7 GB
GLM 4.7 Flash18.2 GB3.0 GB
Hy Memory 30B A3B18.5 GB3.3 GB
Qwen3 Coder 30B A3B Instruct18.7 GB3.5 GB
granite 4.0 h small18.9 GB3.7 GB

Questions

What is the biggest model that runs on the Arc A770 16GB?

Of the models we track, the largest that fits at 8K context is flint (40B) at Q2_K, needing about 14.8 GB.

How much of the Arc A770 16GB's memory can a model use?

About 15.2 GB: we assume 95% of the 16 GB is usable after driver and display overhead.

Spec source: Intel specifications. Model data: Hugging Face Hub API. Speed is a bandwidth ceiling (bandwidth ÷ bytes read per token), not a benchmark. Methodology · Report an error.