RunsOnMy ModelsGPUs & MacsBy VRAMBest GPUCan I run…?OllamaStart hereAPI pricesVRAM calculatorNew

What AI models run on the RX 7600?

The RX 7600 has 8 GB of VRAM (288 GB/s). About 7.6 GB is usable for a model. 161 of 298 open models we track fit at 8K context.

161models fit (8K ctx)
114still fit at 32K ctx
16Blargest that fits
288 GB/smemory bandwidth

This card has 8 GB of VRAM. See every model that fits in 8 GB VRAM. Other 8 GB hardware: RTX 3060 Ti, RTX 3070, RTX 4060, RTX 4060 Ti 8GB, RTX 5060.

Compare for local LLMs: RX 7600 vs RTX 4060 all comparisons

Using a specific tool? Ollama models that run on the RX 7600 · LM Studio · llama.cpp offload

Models that fit, largest first

ModelParamsBest quant (8K)Needs32K contextMax tok/s (est.)Details
Kimi VL A3B Thinking 250616B (3.0B act.)Q2_K6.7 GBQ2_K256details
Kimi VL A3B Thinking16B (3.0B act.)Q2_K6.7 GBQ2_K256details
Kimi VL A3B Instruct16B (3.0B act.)Q2_K7.1 GBno239details
LLaDA2.1 Mini16B (1.4B act.)Q2_K6.7 GBno536details
Moonlight 16B A3B Instruct16B (3.0B act.)Q2_K7.1 GBno233details
DeepSeek Coder V2 Lite Instruct16B (2.7B act.)Q2_K6.3 GBQ2_K289details
DeepSeek V2 Lite16B (2.7B act.)Q2_K6.3 GBQ2_K289details
DeepSeek V2 Lite Chat16B (2.7B act.)Q2_K6.8 GBQ2_K264details
Qwen2.5 Coder 14B Instruct15BQ2_K7.6 GBno50details
DeepSeek R1 Distill Qwen 14B15BQ2_K7.6 GBno50details
Qwen3 14B15BQ2_K7.1 GBno52details
Phi 415BQ2_K7.3 GBno52details
Gemma 3 12B IT12BQ2_K6.2 GBno60details
Llama Guard 4 12B12BQ2_K6.5 GBno64details
KORMo Web 11B11BQ2_K5.8 GBno69details
Llama 3.2 11B Vision Instruct11BQ3_K_M6.8 GBno55details
Llama 3.2 11B Vision11BQ3_K_M6.7 GBno55details
GLM 4.6V Flash10BQ4_K_M6.7 GBQ3_K_M47details
GLM 4.1V 9B Thinking10BQ4_K_M6.7 GBQ3_K_M47details
Step3 VL 10B10BQ4_K_M6.5 GBno57details
SenseNova Loop 9B Preview9.7BQ4_K_M6.4 GBQ3_K_M51details
Qwen3.5 9B9.7BQ3_K_M5.9 GBQ3_K_M56details
Ornith 1.5 9B9.7BQ4_K_M6.6 GBQ3_K_M49details
SenseNova Loop 9B9.4BQ4_K_M6.4 GBQ3_K_M51details
MiMo V2.6 Distill Qwen 9B9.4BQ4_K_M6.6 GBQ3_K_M49details
GLM 4 9B9.4BQ4_K_M6.8 GBQ3_K_M46details
Apertus v1.5 8B8.9BQ4_K_M6.7 GBno53details
InternLM3 8B Instruct8.8BQ4_K_M6.0 GBQ3_K_M54details
Granite 4.2 8B8.8BQ4_K_M6.8 GBno54details
Granite 4.1 8B8.8BQ4_K_M6.8 GBno54details
Granite 4.1 8B Llmon Preview8.8BQ4_K_M6.8 GBno54details
Qwen3 VL 8B Instruct8.8BQ4_K_M6.5 GBno57details
Molmo2 8B8.7BQ4_K_M6.5 GBno57details
LFM2.5 8B A1B8.5B (1.0B act.)Q5_K_M6.4 GBQ5_K_M404details
Qwen2.5 VL 7B Instruct8.3BQ5_K_M6.2 GBQ4_K_M53details
UI TARS 1.5 7B8.3BQ5_K_M6.2 GBQ4_K_M53details
olmOCR 2 7B 10258.3BQ5_K_M6.2 GBQ4_K_M53details
olmOCR 7B 0225 Preview8.3BQ5_K_M6.2 GBQ4_K_M53details
UI TARS 7B DPO8.3BQ5_K_M6.2 GBQ4_K_M53details
Qwen3 8B8.2BQ4_K_M6.5 GBno57details
DeepSeek R1 0528 Qwen3 8B8.2BQ4_K_M6.5 GBno57details
Granite 3.3 8B Instruct8.2BQ4_K_M6.3 GBno58details
Granite 3.1 8B Instruct8.2BQ4_K_M6.3 GBno58details
Granite 3.2 8B Instruct8.2BQ4_K_M6.3 GBno58details
Apertus 8B Instruct 25098.1BQ4_K_M6.4 GBno57details
Llama 3.1 8B Instruct8.0BQ4_K_M6.2 GBno59details
Llama 3.1 8B8.0BQ4_K_M6.2 GBno59details
Hermes 3 Llama 3.1 8B8.0BQ4_K_M6.2 GBno59details
Llama Guard 3 8B8.0BQ4_K_M6.2 GBno59details
DeepSeek R1 Distill Llama 8B8.0BQ4_K_M6.2 GBno59details
Meta Llama 3 8B Instruct8.0BQ4_K_M6.2 GBno59details
Meta Llama Guard 2 8B8.0BQ4_K_M6.2 GBno59details
Meta Llama 3.1 8B Instruct8.0BQ4_K_M6.2 GBno59details
Aya Expanse 8B8.0BQ4_K_M6.6 GBno57details
Ling 3.0 Tiny7.9B (1.4B act.)Q5_K_M6.2 GBQ4_K_M288details
MiMo 7B RL7.8BQ4_K_M6.2 GBno61details
InternLM2.5 7B Chat7.7BQ5_K_M6.7 GBQ2_K52details
InternLM2 Chat 7B7.7BQ5_K_M6.7 GBQ2_K52details
Qwen2.5 7B Instruct7.6BQ5_K_M6.2 GBQ4_K_M53details
DeepSeek R1 Distill Qwen 7B7.6BQ5_K_M6.2 GBQ4_K_M53details
Qwen2.5 Coder 7B Instruct7.6BQ5_K_M6.2 GBQ4_K_M53details
Olmo 3 7B Instruct7.3BQ3_K_M6.5 GBno79details
Olmo 3 1025 7B7.3BQ3_K_M6.4 GBno81details
Olmo 3 7B Think7.3BQ3_K_M6.5 GBno79details
Mistral 7B v0.17.2BQ5_K_M6.2 GBQ2_K56details
Mistral 7B Instruct v0.27.2BQ5_K_M6.2 GBQ2_K56details
Mistral 7B Instruct v0.17.2BQ5_K_M6.2 GBQ2_K56details
Falcon 7B7.2BQ5_K_M5.9 GBQ5_K_M50details
Falcon 7B Instruct7.2BQ5_K_M5.7 GBQ5_K_M52details
Granite 4.0 H Tiny6.9B (1.5B act.)Q6_K6.0 GBQ6_K239details
OLMoE 1B 7B 09246.9B (1.3B act.)Q6_K6.8 GBQ2_K273details
DeepSeek Coder 7B Instruct v1.56.9BQ3_K_M7.6 GBno83details
DeepSeek Coder 6.7B Instruct6.7BQ3_K_M7.5 GBno87details
DeepSeek Coder 6.7B Base6.7BQ3_K_M7.5 GBno87details
Llama 2 7B HF6.7BQ2_K6.8 GBno114details
Yi 6B 200K6.1BQ6_K5.7 GBQ5_K_M58details
Yi 6B6.1BQ6_K5.7 GBQ5_K_M58details
Molmo2 4B4.9BQ8_06.2 GBQ3_K_M61details
WeVisOne VL 4B4.8BQ8_06.2 GBQ3_K_M61details
Mage VL Cowork4.7BQ8_06.2 GBQ3_K_M61details
Qianfan OCR4.7BQ8_06.2 GBQ3_K_M61details
Qwen3.5 4B4.7BQ8_05.4 GBQ8_062details
SingVL S1 4B4.5BQ8_05.3 GBQ8_064details
LightOnOCR 3 4B4.5BQ8_05.3 GBQ8_064details
Intern Decision 4B4.5BQ8_05.3 GBQ8_064details
Qwen3 VL 4B Instruct4.4BQ8_05.8 GBQ2_K67details
Gemma 3 4B IT4.3BQ8_05.0 GBQ8_070details
MedGemma 1.5 4B IT4.3BQ8_05.0 GBQ8_070details
MedGemma 4B IT4.3BQ8_05.0 GBQ8_070details
TranslateGemma 4B IT4.3BQ8_05.0 GBQ8_070details
Gemma 3 4B PT4.3BQ8_05.0 GBQ8_070details
Phi 3.5 Vision Instruct4.1BQ6_K6.6 GBno85details
Qwen3 4B4.0BQ8_05.8 GBQ2_K67details
Qwen3 4B Instruct 25074.0BQ8_05.8 GBQ2_K67details
Granite Vision 4.1 4B4.0BQ8_04.6 GBQ8_080details
Phi 4 Mini Instruct3.8BQ8_05.6 GBQ3_K_M71details
Phi 3 Mini 4K Instruct3.8BQ6_K6.4 GBno92details
Phi 3.5 Mini Instruct3.8BQ6_K6.4 GBno92details
Phi 3 Mini 128K Instruct3.8BQ6_K6.4 GBno92details
Qwen2.5 VL 3B Instruct3.8BQ8_04.0 GBQ8_088details
Granite 4.2 3B3.7BQ8_04.8 GBQ8_074details
Granite 4.1 3B Llmon Preview3.7BQ8_04.8 GBQ8_074details
Raon 1 Nano3.5BQ8_04.3 GBQ8_085details
Granite 4.1 3B3.4BQ8_04.6 GBQ8_080details
Granite 4.0 Micro3.4BQ8_04.6 GBQ8_080details
DeepSeek OCR 23.4B (1.2B act.)F166.6 GBQ8_0139details
DeepSeek VL2 Tiny3.4B (1.2B act.)Q8_04.5 GBQ8_0231details
Tiny Aya Global3.3BQ8_04.6 GBQ8_081details
Unlimited OCR 23.3B (1.1B act.)F166.6 GBQ8_0142details
Unlimited OCR3.3B (1.1B act.)F166.6 GBQ8_0143details
DeepSeek OCR3.3B (1.1B act.)F166.6 GBQ8_0143details
Granite 3.1 3B A800M Instruct3.3B (883M act.)Q8_04.3 GBQ8_0306details
Llama 3.2 3B Instruct3.2BQ8_04.7 GBQ6_K84details
Llama 3.2 3B3.2BQ8_04.7 GBQ6_K84details
Hermes 3 Llama 3.2 3B3.2BQ8_04.7 GBQ6_K84details
d1 3B3.1BF165.8 GBF1653details
LFM2.5 VL 3B3.1BF165.8 GBF1653details
Qwen2.5 3B Instruct3.1BF166.7 GBQ8_047details
SmolLM3 3B3.1BQ8_04.3 GBQ8_088details
SmolLM3 3B Base3.1BQ8_04.3 GBQ8_088details
Phi 22.8BQ8_05.8 GBno97details
LFM2.5 2.6B2.7BQ8_03.4 GBQ8_0100details
Granite 3.3 2B Instruct2.5BF165.8 GBQ8_057details
MiniCPM5 2B2.5BF165.7 GBF1657details
WeVisOne VL 2B2.4BF165.4 GBQ8_071details
Cosmos Reason2 2B2.4BF165.4 GBQ8_071details
Qwen3.5 2B2.3BF164.6 GBF1674details
SmolVLM Instruct2.2BF165.4 GBQ4_K_M79details
Qwen3 1.7B2.0BF165.4 GBQ8_071details
d1 2B1.9BF164.0 GBF1683details
LFM3 VL 2B1.9BF164.0 GBF1683details
DeepSeek R1 Distill Qwen 1.5B1.8BF164.2 GBF1681details
LFM3 1.7B1.7BQ8_02.5 GBQ8_0156details
LFM3 1.7B Base1.7BF164.0 GBF1683details
SmolLM2 1.7B Instruct1.7BF165.2 GBQ4_K_M84details
SmolLM2 1.7B1.7BF165.2 GBQ4_K_M84details
SmolLM 1.7B1.7BF165.2 GBQ4_K_M84details
Granite 4.0 1B1.6BF164.3 GBF1688details
Qwen2.5 1.5B Instruct1.5BF163.8 GBF1693details
OLMo 2 0425 1B1.5BF164.4 GBQ8_097details
DeepSeek Coder 1.3B Instruct1.3BF164.5 GBQ6_K107details
Llama 3.2 1B1.2BF163.2 GBF16116details
Llama 3.2 1B Instruct1.2BF163.2 GBF16116details
LFM2.5 1.2B Instruct1.2BF165.0 GBF1661details
LFM2 1.2B1.2BF163.0 GBF16123details
HunyuanOCR1.1BF163.5 GBF16129details
PaddleOCR VL 1.6959MF162.5 GBF16150details
Qwen3.5 0.8B873MF162.4 GBF16185details
LightOnOCR 3 0.8B853MF162.4 GBF16190details
FastVLM 0.5B759MF162.2 GBF16190details
Qwen3 0.6B752MF163.0 GBF16191details
GOT OCR2.0716MF162.8 GBF16201details
Nemotron 3.5 Parse631MF161.9 GBF16220details
GOT OCR 2.0 HF561MF162.5 GBF16257details
SmolVLM2 500M Video Instruct507MF161.6 GBF16351details
SmolVLM 500M Instruct507MF161.6 GBF16351details
Qwen2.5 0.5B Instruct494MF161.7 GBF16291details
LFM2.5 VL 450M449MF161.3 GBF16405details
SmolLM2 360M Instruct362MF161.5 GBF16397details
SmolLM2 360M362MF161.5 GBF16397details
LFM2.5 350M354MF161.9 GBF16202details

Bigger models with CPU offload

These popular models don't fit in the RX 7600's VRAM at 4-bit, but run with llama.cpp offload, keeping part of the model in system RAM. Speed range is DDR4-3200 to DDR5-6000 (estimate). Method.

ModelParamsOffload plan
Qwen3.8 27B28BOffload Q4_K_M, 32 GB RAM · ~1.9–3.8 tok/s
GLM 5.3 Flash321B MoEOffload Q2_K, 128 GB RAM · ~6.4–12 tok/s
Gemma 4 31B IT31BOffload Q4_K_M, 32 GB RAM · ~1.6–3.2 tok/s
gpt-oss 20B21B MoEOffload Q4_K_M, 32 GB RAM · ~27–45 tok/s
Gemma 4 26B A4B IT26B MoEOffload Q4_K_M, 32 GB RAM · ~24–39 tok/s
Qwen3.6 35B A3B36B MoEOffload Q4_K_M, 32 GB RAM · ~30–44 tok/s
Qwen3 32B33BOffload Q4_K_M, 32 GB RAM · ~1.6–3.2 tok/s
Qwen3.6 27B28BOffload Q4_K_M, 32 GB RAM · ~1.9–3.7 tok/s
gpt-oss 120B117B MoEOffload Q4_K_M, 64 GB RAM · ~12–23 tok/s
DeepSeek V4 Flash 0731304B MoEOffload Q2_K, 128 GB RAM · ~2.3–4.7 tok/s
Qwen3 Coder 30B A3B Instruct31B MoEOffload Q4_K_M, 32 GB RAM · ~24–44 tok/s
Qwen3.5 27B28BOffload Q4_K_M, 32 GB RAM · ~1.9–3.7 tok/s
Qwen3.5 35B A3B36B MoEOffload Q4_K_M, 32 GB RAM · ~30–44 tok/s
DeepSeek V4 Flash291B MoEOffload Q2_K, 128 GB RAM · ~8.3–15 tok/s
GLM 4.5 Air110B MoEOffload Q4_K_M, 128 GB RAM · ~2.5–5.2 tok/s
MiniMax M2.5229B MoEOffload Q3_K_M, 128 GB RAM · ~6.3–12 tok/s
Hy3299B MoEOffload Q2_K, 128 GB RAM · ~4.9–9.5 tok/s

Comparing cards? The RX 7600 (launch MSRP $269) is in best GPUs for local LLMs under $300. See also Mac vs PC.

Just too big at 4-bit

These need up to 35% more memory than the RX 7600 has at Q4_K_M and 8K context. A smaller quant or shorter context may work.

ModelQ4_K_M needsShort by
Llama 3.2 11B Vision8.0 GB0.4 GB
KORMo Web 11B8.3 GB0.7 GB
DeepSeek Coder 6.7B Instruct8.3 GB0.7 GB
DeepSeek Coder 6.7B Base8.3 GB0.7 GB
DeepSeek Coder 7B Instruct v1.58.3 GB0.7 GB
Llama 2 7B HF8.3 GB0.7 GB
Gemma 3 12B IT8.5 GB0.9 GB
Llama Guard 4 12B9.1 GB1.5 GB
DeepSeek V2 Lite9.8 GB2.2 GB

Questions

What is the biggest model that runs on the RX 7600?

Of the models we track, the largest that fits at 8K context is Kimi VL A3B Thinking 2506 (16B) at Q2_K, needing about 6.7 GB.

How much of the RX 7600's memory can a model use?

About 7.6 GB: we assume 95% of the 8 GB is usable after driver and display overhead.

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