Radeon PRO W7800 (32GB)
The Radeon PRO W7800 (32GB) provides 32 GB of memory with 576 GB/s bandwidth, supporting local inference via ROCm and Vulkan acceleration.
Memory
32 GB
Usable Memory Ceiling
32 GB
Bandwidth
576 GB/s
Max Inference Class
32B Models (Q4)
TDP
260W
VRAM Scaling by Context Length
Estimated memory requirements as context length increases. Curves crossing above the reference line indicate Out of Memory (OOM).
Precision:
Best Models to Run on Radeon PRO W7800 (32GB)
Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on Radeon PRO W7800 (32GB).
Sort by:
Context:
KV Cache:
GPUs:
| Rank | Model | Parameters | Precision | TPS (Approx.) | TTFT | Can Run |
|---|---|---|---|---|---|---|
#48 | Qwen 3.8 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
#83 | K2 Horizon MoVA 36B A4B | 36B (4B active) | INT4 | 104 tok/s | ~737 ms | |
#93 | Qwen3.5-27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
#105 | Sarvam-30B | 32B (2.4B active) | INT4 | 129 tok/s | ~516 ms | |
#106 | Qwen3.6 35B A3B | 35B (3B active) | INT4 | 118 tok/s | ~601 ms | |
#110 | Muse Glimmer 30B | 30B | INT4 Q8 | 26 tok/s 14 tok/s | ~4.1s ~4.9s | |
#111 | MiMo V2 Flash | 15B (309B active) | INT4 Q8 | 3 tok/s 2 tok/s | ~40.8s ~41.1s | |
#112 | Phi-4 Reasoning Plus | 14B | INT4 Q8 FP16 | 37 tok/s 25 tok/s 13 tok/s | ~2.0s ~2.0s ~2.4s | |
#117 | Gemma 4 26B A4B | 25.2B (3.8B active) | INT4 Q8 | 110 tok/s 76 tok/s | ~675 ms ~736 ms | |
#125 | Qwen3.5-9B | 9B | INT4 Q8 FP16 | 70 tok/s 44 tok/s 24 tok/s | ~1.3s ~1.3s ~1.3s | |
#126 | Qwen3.5-35B-A3B | 35B (3B active) | INT4 | 118 tok/s | ~601 ms | |
#128 | K2 Horizon 7B | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
#135 | Gemma 4 12B | 11.95B | INT4 Q8 FP16 | 42 tok/s 29 tok/s 16 tok/s | ~1.7s ~1.7s ~1.9s | |
#140 | Qwen3.5-4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
#143 | GPT-OSS 20B | 21B (3.6B active) | INT4 Q8 | 57 tok/s 47 tok/s | ~640 ms ~697 ms | |
#144 | GLM-4V | 9B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
#150 | NVIDIA Nemotron 3 Nano 30B-A3B | 3.5B (30B active) | INT4 Q8 FP16 | 25 tok/s 15 tok/s 8 tok/s | ~4.0s ~4.0s ~4.1s | |
#153 | Qwen3-30B-A3B | 30B (3B active) | INT4 Q8 | 119 tok/s 81 tok/s | ~589 ms ~694 ms | |
#158 | Granite 4.2 8B | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#159 | Gemma 4 E4B | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#163 | Mistral Large 3 | 41B (675B active) | INT4 | 1 tok/s | ~104.7s | |
#165 | GLM-4 | 32B | INT4 | 19 tok/s | ~4.4s | |
#166 | Phi-4 | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
#167 | Gemma 3 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
#170 | DeepSeek-R1 32B | 32B | INT4 Q8 | 25 tok/s 13 tok/s | ~4.4s ~5.2s | |
#171 | Qwen3-32B | 32B | INT4 | 25 tok/s | ~4.4s | |
#175 | Magistral Small | 24B | INT4 Q8 | 24 tok/s 15 tok/s | ~3.3s ~3.7s | |
#176 | OLMo 3.1 32B Instruct | 32B | INT4 | 19 tok/s | ~4.4s | |
#177 | Qwen2.5-32B | 32B | INT4 | 25 tok/s | ~4.4s | |
#178 | OLMo 3 7B Think | 7B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
#179 | DeepSeek-R1 14B | 14B | INT4 Q8 FP16 | 51 tok/s 31 tok/s 16 tok/s | ~2.0s ~2.0s ~2.2s | |
#183 | Ministral 3 14B | 14B | INT4 Q8 FP16 | 37 tok/s 25 tok/s 13 tok/s | ~2.0s ~2.0s ~2.4s | |
#186 | Gemma 3 12B | 12B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
#188 | Mistral-Small-2501 | 24B | INT4 Q8 | 32 tok/s 18 tok/s | ~3.3s ~3.6s | |
#189 | Ministral 3 8B | 8B | INT4 Q8 FP16 | 57 tok/s 41 tok/s 24 tok/s | ~1.1s ~1.1s ~1.2s | |
#190 | Gemma 4 31B | 30.7B | INT4 Q8 | 26 tok/s 13 tok/s | ~4.2s ~5.0s | |
#191 | MiniCPM5-2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
#192 | Qwen3.5-2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
#193 | Gemma 2 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
#194 | Qwen2.5-14B | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
#195 | OLMo 3.1 32B Think | 32B | INT4 | 19 tok/s | ~4.4s | |
#196 | Nemotron 3.5 Lightning | 30B (3B active) | INT4 | 52 tok/s | ~600 ms | |
#200 | Gemma 4 E2B | 5.1B | INT4 Q8 FP16 | 103 tok/s 70 tok/s 40 tok/s | ~741 ms ~746 ms ~756 ms | |
#202 | DeepSeek-R1 8B | 8B | INT4 Q8 FP16 | 79 tok/s 50 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#203 | Kimi Linear 48B A3B Instruct | 48B (3B active) | INT4 | 40 tok/s | ~761 ms | |
#204 | Qwen3-4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
#205 | Ministral 3 3B | 3B | INT4 Q8 FP16 | 99 tok/s 79 tok/s 53 tok/s | ~456 ms ~458 ms ~464 ms | |
#206 | OLMo 3 32B Base | 32B | INT4 | 19 tok/s | ~4.4s | |
#207 | Gemma 2 9B | 9B | INT4 Q8 FP16 | 70 tok/s 44 tok/s 24 tok/s | ~1.3s ~1.3s ~1.3s | |
#208 | Qwen3.5-0.8B | 0.8B | INT4 Q8 FP16 | 220 tok/s 189 tok/s 145 tok/s | ~133 ms ~134 ms ~136 ms | |
#209 | Qwen3-14B | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
#210 | Gemma 3 4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
#211 | Qwen3-8B | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#213 | DeepSeek-R1 7B | 7B | INT4 Q8 FP16 | 86 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
#214 | Qwen3-1.7B | 1.7B | INT4 Q8 FP16 | 175 tok/s 137 tok/s 93 tok/s | ~272 ms ~274 ms ~277 ms | |
#215 | Llama 3.1 8B | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#216 | Gemma 3 270M | 0.27B | INT4 Q8 FP16 | 244 tok/s 231 tok/s 205 tok/s | ~50 ms ~50 ms ~51 ms | |
#217 | Command R | 35B | INT4 | 18 tok/s | ~4.8s | |
#218 | Qwen3-0.6B | 0.6B | INT4 Q8 FP16 | 234 tok/s 207 tok/s 165 tok/s | ~101 ms ~101 ms ~103 ms | |
#219 | Llama 3 8B | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#221 | Phi-4-Mini | 3.8B | INT4 Q8 FP16 | 122 tok/s 86 tok/s 51 tok/s | ~563 ms ~566 ms ~574 ms | |
#222 | Mixtral-8x7B-v0.1 | 46.7B (7B active) | INT4 | 72 tok/s | ~1.3s | |
#223 | Qwen2.5-7B | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
#224 | Ministral-8B-2410 | 8B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
#225 | DeepSeek-R1 1.5B | 1.5B | INT4 Q8 FP16 | 183 tok/s 146 tok/s 101 tok/s | ~242 ms ~243 ms ~246 ms | |
#226 | Llama 3.2 3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
#227 | Nemotron 3.5 Lightning 30B A3B | 30B (3B active) | INT4 Q8 | 119 tok/s 81 tok/s | ~589 ms ~694 ms | |
#228 | Phi-3-mini | 3.8B | INT4 Q8 FP16 | 122 tok/s 86 tok/s 51 tok/s | ~563 ms ~566 ms ~574 ms | |
#229 | Gemma 3n E2B IT | 6B (2B active) | INT4 Q8 FP16 | 99 tok/s 86 tok/s 64 tok/s | ~342 ms ~343 ms ~347 ms | |
#230 | Mistral-7B-Instruct-v0.1 | 7.3B | INT4 Q8 FP16 | 82 tok/s 53 tok/s 30 tok/s | ~1.0s ~1.0s ~1.1s | |
#231 | Llama 3.2 1B | 1B | INT4 Q8 FP16 | 207 tok/s 174 tok/s 128 tok/s | ~168 ms ~168 ms ~171 ms | |
#232 | Gemma 2 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
#233 | Phi-3-medium | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
#234 | OLMo 3 7B Instruct | 7B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
#235 | Yi-34B | 34B | INT4 | 18 tok/s | ~4.7s | |
#236 | Qwen2-7B | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
#237 | Phi-3-small | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
#238 | Gemma 3 1B | 1B | INT4 Q8 FP16 | 207 tok/s 174 tok/s 128 tok/s | ~168 ms ~168 ms ~171 ms | |
#239 | Mistral-7B-Instruct-v0.2 | 7.3B | INT4 Q8 FP16 | 82 tok/s 53 tok/s 30 tok/s | ~1.0s ~1.0s ~1.1s | |
#241 | Gemma 1 7B | 7B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
#242 | Gemma 1 2B | 2B | INT4 Q8 FP16 | 172 tok/s 130 tok/s 85 tok/s | ~316 ms ~318 ms ~322 ms | |
#243 | ChatGLM3-6B | 6B | INT4 Q8 FP16 | 69 tok/s 51 tok/s 31 tok/s | ~868 ms ~873 ms ~886 ms | |
#244 | ChatGLM2-6B | 6B | INT4 Q8 FP16 | 69 tok/s 51 tok/s 31 tok/s | ~868 ms ~873 ms ~886 ms | |
#245 | ChatGLM-6B | 6B | INT4 Q8 FP16 | 69 tok/s 51 tok/s 31 tok/s | ~868 ms ~873 ms ~886 ms | |
Unranked Models (Sorted by Speed) | ||||||
- | DINOv3 ViT-L/16 | 0.3B | INT4 Q8 FP16 | 258 tok/s 241 tok/s 210 tok/s | ~54 ms ~55 ms ~55 ms | |
- | Qwen2-0.5B | 0.5B | INT4 Q8 FP16 | 242 tok/s 218 tok/s 178 tok/s | ~85 ms ~85 ms ~86 ms | |
- | Qwen2.5-0.5B | 0.5B | INT4 Q8 FP16 | 242 tok/s 218 tok/s 178 tok/s | ~85 ms ~85 ms ~86 ms | |
- | ERNIE-4.5-0.3B | 0.3B | INT4 Q8 FP16 | 239 tok/s 225 tok/s 198 tok/s | ~55 ms ~55 ms ~55 ms | |
- | ERNIE-4.5-0.3B-Base | 0.3B | INT4 Q8 FP16 | 239 tok/s 225 tok/s 198 tok/s | ~55 ms ~55 ms ~55 ms | |
- | Liquid D1-Omni 600M | 0.6B | INT4 Q8 FP16 | 234 tok/s 207 tok/s 165 tok/s | ~101 ms ~101 ms ~103 ms | |
- | PixArt-α XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 234 tok/s 206 tok/s 164 tok/s | ~102 ms ~103 ms ~104 ms | |
- | PixArt-Σ XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 234 tok/s 206 tok/s 164 tok/s | ~102 ms ~103 ms ~104 ms | |
- | Nemotron 3.5 Parse | 0.63B | INT4 Q8 FP16 | 232 tok/s 204 tok/s 162 tok/s | ~106 ms ~107 ms ~108 ms | |
- | SD 2.1 Turbo | 0.86B | INT4 Q8 FP16 | 216 tok/s 185 tok/s 139 tok/s | ~143 ms ~144 ms ~146 ms | |
- | Stable Diffusion 1.4 | 0.86B | INT4 Q8 FP16 | 216 tok/s 185 tok/s 139 tok/s | ~143 ms ~144 ms ~146 ms | |
- | Stable Diffusion 2.1 | 0.86B | INT4 Q8 FP16 | 216 tok/s 185 tok/s 139 tok/s | ~143 ms ~144 ms ~146 ms | |
- | Falcon-1B | 1B | INT4 Q8 FP16 | 214 tok/s 179 tok/s 131 tok/s | ~167 ms ~168 ms ~170 ms | |
- | Falcon3-1B | 1B | INT4 Q8 FP16 | 207 tok/s 174 tok/s 128 tok/s | ~168 ms ~168 ms ~171 ms | |
- | MiniCPM-V 4.7 1B | 1B | INT4 Q8 FP16 | 207 tok/s 174 tok/s 128 tok/s | ~168 ms ~168 ms ~171 ms | |
- | Kandinsky 2.2 Decoder | 1.22B | INT4 Q8 FP16 | 196 tok/s 161 tok/s 115 tok/s | ~200 ms ~201 ms ~204 ms | |
- | Wan 2.1 T2V 1.3B | 1.3B | INT4 Q8 FP16 | 192 tok/s 156 tok/s 110 tok/s | ~211 ms ~213 ms ~215 ms | |
- | Qwen2-1.5B | 1.5B | INT4 Q8 FP16 | 183 tok/s 146 tok/s 101 tok/s | ~242 ms ~243 ms ~246 ms | |
- | Qwen2.5-1.5B | 1.5B | INT4 Q8 FP16 | 183 tok/s 146 tok/s 101 tok/s | ~242 ms ~243 ms ~246 ms | |
- | SVD Image-to-Video XT | 1.5B | INT4 Q8 FP16 | 183 tok/s 146 tok/s 101 tok/s | ~242 ms ~243 ms ~246 ms | |
- | SVD Image-to-Video XT 1.1 | 1.5B | INT4 Q8 FP16 | 183 tok/s 146 tok/s 101 tok/s | ~242 ms ~243 ms ~246 ms | |
- | LFM3 1.7B | 1.7B | INT4 Q8 FP16 | 175 tok/s 137 tok/s 93 tok/s | ~272 ms ~274 ms ~277 ms | |
- | LFM3 1.7B Base | 1.7B | INT4 Q8 FP16 | 175 tok/s 137 tok/s 93 tok/s | ~272 ms ~274 ms ~277 ms | |
- | Spark X2.5 1.7B | 1.7B | INT4 Q8 FP16 | 175 tok/s 137 tok/s 93 tok/s | ~272 ms ~274 ms ~277 ms | |
- | Hy-MT2 1.8B | 1.8B | INT4 Q8 FP16 | 171 tok/s 134 tok/s 90 tok/s | ~286 ms ~288 ms ~291 ms | |
- | CogVideoX 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | LFM3 Audio 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | LFM3 Omni 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | LFM3 VL 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | Liquid D1 2B | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | LTX-Video 0.9.1 | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | Moondream 3.1 | 9B (2B active) | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~356 ms ~358 ms ~362 ms | |
- | Stable Diffusion 3 Medium | 2B | INT4 Q8 FP16 | 164 tok/s 126 tok/s 83 tok/s | ~317 ms ~319 ms ~323 ms | |
- | Kandinsky 3.0 | 2.1B | INT4 Q8 FP16 | 161 tok/s 123 tok/s 80 tok/s | ~330 ms ~332 ms ~337 ms | |
- | LTX-2 | 2.2B | INT4 Q8 FP16 | 158 tok/s 120 tok/s 78 tok/s | ~344 ms ~346 ms ~350 ms | |
- | SDXL 1.0 Refiner | 2.31B | INT4 Q8 FP16 | 155 tok/s 117 tok/s 75 tok/s | ~359 ms ~361 ms ~366 ms | |
- | Stable Diffusion 3.5 Medium | 2.5B | INT4 Q8 FP16 | 150 tok/s 112 tok/s 71 tok/s | ~385 ms ~387 ms ~392 ms | |
- | SDXL 1.0 Base | 2.56B | INT4 Q8 FP16 | 148 tok/s 110 tok/s 70 tok/s | ~393 ms ~395 ms ~401 ms | |
- | SDXL Turbo | 2.56B | INT4 Q8 FP16 | 148 tok/s 110 tok/s 70 tok/s | ~393 ms ~395 ms ~401 ms | |
- | Falcon-3B | 3B | INT4 Q8 FP16 | 144 tok/s 103 tok/s 63 tok/s | ~452 ms ~455 ms ~461 ms | |
- | CroissantLLM Base | 1.3B | INT4 Q8 FP16 | 142 tok/s 124 tok/s 93 tok/s | ~213 ms ~214 ms ~217 ms | |
- | Phi-1 | 1.3B | INT4 Q8 FP16 | 142 tok/s 124 tok/s 93 tok/s | ~213 ms ~214 ms ~217 ms | |
- | Phi-1.5 | 1.3B | INT4 Q8 FP16 | 142 tok/s 124 tok/s 93 tok/s | ~213 ms ~214 ms ~217 ms | |
- | DeepSeek-R1 3B | 3B | INT4 Q8 FP16 | 140 tok/s 101 tok/s 62 tok/s | ~453 ms ~455 ms ~462 ms | |
- | Mellum 2.1 12B A2.5B Thinking | 12B (2.5B active) | INT4 Q8 FP16 | 140 tok/s 106 tok/s 69 tok/s | ~442 ms ~445 ms ~450 ms | |
- | Falcon3-3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | LFM2-VL 3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | Liquid D1 3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | Liquid D1-VL 3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | Liquid LFM2-VL 3B D1 | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | Qwen2.5-3B | 3B | INT4 Q8 FP16 | 138 tok/s 100 tok/s 62 tok/s | ~453 ms ~456 ms ~462 ms | |
- | LFM2.5 24B A2B | 24B (2.5B active) | INT4 Q8 | 131 tok/s 96 tok/s | ~501 ms ~546 ms | |
- | ERNIE-4.5-21B-A3B | 21B (3B active) | INT4 Q8 | 124 tok/s 93 tok/s | ~552 ms ~555 ms | |
- | ERNIE-4.5-21B-A3B-Base | 21B (3B active) | INT4 Q8 | 124 tok/s 93 tok/s | ~552 ms ~555 ms | |
- | ERNIE-4.5-VL-28B-A3B | 28B (3B active) | INT4 Q8 | 120 tok/s 86 tok/s | ~580 ms ~633 ms | |
- | ERNIE-4.5-VL-28B-A3B-Base | 28B (3B active) | INT4 Q8 | 120 tok/s 86 tok/s | ~580 ms ~633 ms | |
- | FLUX.2 Klein 4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
- | FrogNano 4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
- | Hy-Memory 30B A3B | 30B (3B active) | INT4 Q8 | 119 tok/s 81 tok/s | ~589 ms ~694 ms | |
- | Hy-MT2 30B A3B | 30B (3B active) | INT4 Q8 | 119 tok/s 81 tok/s | ~589 ms ~694 ms | |
- | Mage-VL Cowork | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
- | Spark X2.5 4B | 4B | INT4 Q8 FP16 | 119 tok/s 83 tok/s 49 tok/s | ~590 ms ~594 ms ~602 ms | |
- | MiniCPM-V 4.7 35B A3B | 35B (3B active) | INT4 | 118 tok/s | ~601 ms | |
- | EmbeddingGemma 2 | 2B | INT4 Q8 FP16 | 115 tok/s 97 tok/s 69 tok/s | ~319 ms ~321 ms ~325 ms | |
- | CogVideoX 1.5 5B | 5B | INT4 Q8 FP16 | 105 tok/s 71 tok/s 41 tok/s | ~727 ms ~732 ms ~742 ms | |
- | CogVideoX 5B | 5B | INT4 Q8 FP16 | 105 tok/s 71 tok/s 41 tok/s | ~727 ms ~732 ms ~742 ms | |
- | Wan 2.2 TI2V 5B | 5B | INT4 Q8 FP16 | 105 tok/s 71 tok/s 41 tok/s | ~727 ms ~732 ms ~742 ms | |
- | Phi-2 | 2.7B | INT4 Q8 FP16 | 103 tok/s 84 tok/s 57 tok/s | ~415 ms ~417 ms ~423 ms | |
- | MaLLaM-3B | 3B | INT4 Q8 FP16 | 99 tok/s 79 tok/s 53 tok/s | ~456 ms ~458 ms ~464 ms | |
- | SmolLM3 3B | 3B | INT4 Q8 FP16 | 99 tok/s 79 tok/s 53 tok/s | ~456 ms ~458 ms ~464 ms | |
- | Falcon-7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
- | AuraFlow v0.2 | 6.8B | INT4 Q8 FP16 | 86 tok/s 56 tok/s 31 tok/s | ~974 ms ~980 ms ~994 ms | |
- | AuraFlow v0.3 | 6.8B | INT4 Q8 FP16 | 86 tok/s 56 tok/s 31 tok/s | ~974 ms ~980 ms ~994 ms | |
- | Falcon3-7B | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
- | Hy-MT2 7B | 7B | INT4 Q8 FP16 | 84 tok/s 55 tok/s 31 tok/s | ~1.0s ~1.0s ~1.0s | |
- | Mistral-7B-v0.1 | 7.3B | INT4 Q8 FP16 | 82 tok/s 53 tok/s 30 tok/s | ~1.0s ~1.0s ~1.1s | |
- | Youtu Parsing Omni | 8B | INT4 Q8 FP16 | 79 tok/s 50 tok/s 27 tok/s | ~1.1s ~1.1s ~1.2s | |
- | Stable Diffusion 3.5 Large | 8.1B | INT4 Q8 FP16 | 76 tok/s 48 tok/s 27 tok/s | ~1.2s ~1.2s ~1.2s | |
- | Stable Diffusion 3.5 Large Turbo | 8.1B | INT4 Q8 FP16 | 76 tok/s 48 tok/s 27 tok/s | ~1.2s ~1.2s ~1.2s | |
- | FLUX.2 Klein 9B | 9B | INT4 Q8 FP16 | 70 tok/s 44 tok/s 24 tok/s | ~1.3s ~1.3s ~1.3s | |
- | LensVLM 9B | 9B | INT4 Q8 FP16 | 70 tok/s 44 tok/s 24 tok/s | ~1.3s ~1.3s ~1.3s | |
- | MiMo V2.6 Distill Qwen 9B | 9B | INT4 Q8 FP16 | 70 tok/s 44 tok/s 24 tok/s | ~1.3s ~1.3s ~1.3s | |
- | ChatGLM3-6B-32K | 6B | INT4 Q8 FP16 | 69 tok/s 51 tok/s 31 tok/s | ~868 ms ~873 ms ~886 ms | |
- | Yi-6B | 6B | INT4 Q8 FP16 | 69 tok/s 51 tok/s 31 tok/s | ~868 ms ~873 ms ~886 ms | |
- | Falcon3-10B | 10B | INT4 Q8 FP16 | 65 tok/s 41 tok/s 22 tok/s | ~1.4s ~1.4s ~1.4s | |
- | Kimi-VL-A3B-Instruct | 16B (3B active) | INT4 Q8 | 65 tok/s 57 tok/s | ~537 ms ~539 ms | |
- | Mochi 1 Preview | 10B | INT4 Q8 FP16 | 65 tok/s 41 tok/s 22 tok/s | ~1.4s ~1.4s ~1.4s | |
- | Falcon2-11B | 11B | INT4 Q8 FP16 | 63 tok/s 38 tok/s 21 tok/s | ~1.6s ~1.6s ~1.6s | |
- | Hunyuan Lite | 7B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
- | OLMo 3 7B Base | 7B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
- | SEA-LION-7B | 7.1B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
- | SEA-LION-7B-Instruct | 7.1B | INT4 Q8 FP16 | 62 tok/s 45 tok/s 27 tok/s | ~1.0s ~1.0s ~1.0s | |
- | FLUX.1 Canny Dev | 11.9B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
- | FLUX.1 Depth Dev | 11.9B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
- | FLUX.1 Dev | 11.9B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
- | FLUX.1 Fill Dev | 11.9B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
- | FLUX.1 Schnell | 11.9B | INT4 Q8 FP16 | 57 tok/s 35 tok/s 19 tok/s | ~1.7s ~1.7s ~1.7s | |
- | Sahabat-AI-Llama3-8B-Instruct | 8B | INT4 Q8 FP16 | 57 tok/s 41 tok/s 24 tok/s | ~1.1s ~1.1s ~1.2s | |
- | Typhoon-2-8B | 8B | INT4 Q8 FP16 | 57 tok/s 41 tok/s 24 tok/s | ~1.1s ~1.1s ~1.2s | |
- | HunyuanVideo 1.5 720P T2V | 13B | INT4 Q8 FP16 | 53 tok/s 32 tok/s 17 tok/s | ~1.8s ~1.8s ~2.0s | |
- | GLM-4-9B | 9B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | GLM-4-9B-Chat | 9B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | GLM-4-9B-Chat-1M | 9B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Sahabat-AI-Gemma2-9B | 9.2B | INT4 Q8 FP16 | 52 tok/s 36 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Sahabat-AI-Gemma2-9B-Instruct | 9.2B | INT4 Q8 FP16 | 52 tok/s 36 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Yi-9B | 9B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Wan 2.1 I2V 14B 720P | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
- | Wan 2.1 T2V 14B | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
- | Wan 2.2 T2V A14B | 14B | INT4 Q8 FP16 | 50 tok/s 30 tok/s 15 tok/s | ~2.0s ~2.0s ~2.2s | |
- | FLUX.2 Dev | 15B | INT4 Q8 FP16 | 47 tok/s 29 tok/s 14 tok/s | ~2.1s ~2.1s ~2.5s | |
- | GreenMind-14B-R1 | 14B | INT4 Q8 FP16 | 37 tok/s 25 tok/s 13 tok/s | ~2.0s ~2.0s ~2.4s | |
- | Arete 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
- | AREX-2 | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
- | PPLX Decider v1 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
- | PPLX Decider v1.1 27B | 27B | INT4 Q8 | 29 tok/s 16 tok/s | ~3.7s ~4.1s | |
- | Agnes 3.0 Flash | 33B | INT4 | 24 tok/s | ~4.5s | |
- | Falcon-40B | 40B | INT4 | 21 tok/s | ~5.5s | |
Inference Capacity
Maximum model parameter class runnable by weight precision.
Context:
4-bit
32B Models
8-bit
14B Models
16-bit
14B Models
Fine-Tuning Capacity
Hardware capacity for local training and adapter fine-tuning.
Context Length:
QLoRA
32B Models
LoRA
14B Models
Full Parameter Training
3B Models
Workload Recommendations
Guidance for optimal precision and operational ceilings on Radeon PRO W7800 (32GB).
Inference Sweet Spot
Optimized for 14B models at uncompressed or Q8 precision, and 32B models at Q4 with up to 32k context.
Bandwidth & Speed Profile
With 576 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 128 tok/s on an 8B Q4 model and 15 tok/s on a 70B Q4 model.
Frequently Asked Questions
©2025 ApX Machine Learning