Radeon PRO W7900 (48GB)
The Radeon PRO W7900 (48GB) provides 48 GB of memory with 864 GB/s bandwidth, supporting local inference via ROCm and Vulkan acceleration.
Memory
48 GB
Usable Memory Ceiling
48 GB
Bandwidth
864 GB/s
Max Inference Class
32B Models (Q4)
TDP
295W
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 W7900 (48GB)
Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on Radeon PRO W7900 (48GB).
Sort by:
Context:
KV Cache:
GPUs:
| Rank | Model | Parameters | Precision | TPS (Approx.) | TTFT | Can Run |
|---|---|---|---|---|---|---|
#48 | Qwen 3.8 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
#83 | K2 Horizon MoVA 36B A4B | 36B (4B active) | INT4 Q8 | 132 tok/s 100 tok/s | ~555 ms ~558 ms | |
#93 | Qwen3.5-27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
#105 | Sarvam-30B | 32B (2.4B active) | INT4 Q8 | 157 tok/s 129 tok/s | ~389 ms ~391 ms | |
#106 | Qwen3.6 35B A3B | 35B (3B active) | INT4 Q8 | 146 tok/s 116 tok/s | ~453 ms ~455 ms | |
#107 | Nemotron 3.5 Bolt 74B A9B | 74B | INT4 | 16 tok/s | ~8.3s | |
#110 | Muse Glimmer 30B | 30B | INT4 Q8 | 37 tok/s 22 tok/s | ~3.1s ~3.1s | |
#111 | MiMo V2 Flash | 15B (309B active) | INT4 Q8 FP16 | 4 tok/s 2 tok/s 1 tok/s | ~30.6s ~30.8s ~31.2s | |
#112 | Phi-4 Reasoning Plus | 14B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.5s ~1.5s ~1.5s | |
#117 | Gemma 4 26B A4B | 25.2B (3.8B active) | INT4 Q8 | 138 tok/s 105 tok/s | ~508 ms ~511 ms | |
#125 | Qwen3.5-9B | 9B | INT4 Q8 FP16 | 94 tok/s 62 tok/s 35 tok/s | ~961 ms ~966 ms ~979 ms | |
#126 | Qwen3.5-35B-A3B | 35B (3B active) | INT4 Q8 | 146 tok/s 116 tok/s | ~453 ms ~455 ms | |
#128 | K2 Horizon 7B | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
#135 | Gemma 4 12B | 11.95B | INT4 Q8 FP16 | 59 tok/s 42 tok/s 25 tok/s | ~1.3s ~1.3s ~1.3s | |
#140 | Qwen3.5-4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
#143 | GPT-OSS 20B | 21B (3.6B active) | INT4 Q8 FP16 | 78 tok/s 69 tok/s 46 tok/s | ~481 ms ~483 ms ~572 ms | |
#144 | GLM-4V | 9B | INT4 Q8 FP16 | 72 tok/s 52 tok/s 32 tok/s | ~964 ms ~969 ms ~982 ms | |
#150 | NVIDIA Nemotron 3 Nano 30B-A3B | 3.5B (30B active) | INT4 Q8 FP16 | 36 tok/s 22 tok/s 11 tok/s | ~3.0s ~3.0s ~3.1s | |
#153 | Qwen3-30B-A3B | 30B (3B active) | INT4 Q8 | 148 tok/s 117 tok/s | ~444 ms ~446 ms | |
#158 | Granite 4.2 8B | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#159 | Gemma 4 E4B | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#162 | Qwen2.5-72B | 72B | INT4 | 16 tok/s | ~8.0s | |
#163 | Mistral Large 3 | 41B (675B active) | INT4 | 2 tok/s | ~66.8s | |
#164 | DeepSeek-R1 70B | 70B | INT4 | 18 tok/s | ~7.2s | |
#165 | GLM-4 | 32B | INT4 Q8 | 28 tok/s 17 tok/s | ~3.3s ~3.6s | |
#166 | Phi-4 | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
#167 | Gemma 3 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
#168 | Llama 3.3 70B | 70B | INT4 | 17 tok/s | ~7.2s | |
#170 | DeepSeek-R1 32B | 32B | INT4 Q8 | 36 tok/s 21 tok/s | ~3.3s ~3.3s | |
#171 | Qwen3-32B | 32B | INT4 Q8 | 35 tok/s 21 tok/s | ~3.3s ~3.3s | |
#173 | Hunyuan Standard | 52B (389B active) | INT4 | 3 tok/s | ~38.6s | |
#175 | Magistral Small | 24B | INT4 Q8 | 35 tok/s 23 tok/s | ~2.5s ~2.5s | |
#176 | OLMo 3.1 32B Instruct | 32B | INT4 Q8 | 28 tok/s 17 tok/s | ~3.3s ~3.6s | |
#177 | Qwen2.5-32B | 32B | INT4 Q8 | 35 tok/s 21 tok/s | ~3.3s ~3.3s | |
#178 | OLMo 3 7B Think | 7B | INT4 Q8 FP16 | 84 tok/s 63 tok/s 39 tok/s | ~756 ms ~761 ms ~770 ms | |
#179 | DeepSeek-R1 14B | 14B | INT4 Q8 FP16 | 71 tok/s 44 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
#181 | Llama 3.1 70B | 70B | INT4 | 17 tok/s | ~7.2s | |
#183 | Ministral 3 14B | 14B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.5s ~1.5s ~1.5s | |
#186 | Gemma 3 12B | 12B | INT4 Q8 FP16 | 77 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
#187 | Qwen2-72B | 72B | INT4 | 16 tok/s | ~8.0s | |
#188 | Mistral-Small-2501 | 24B | INT4 Q8 FP16 | 45 tok/s 27 tok/s 13 tok/s | ~2.5s ~2.5s ~3.0s | |
#189 | Ministral 3 8B | 8B | INT4 Q8 FP16 | 78 tok/s 57 tok/s 35 tok/s | ~859 ms ~864 ms ~875 ms | |
#190 | Gemma 4 31B | 30.7B | INT4 Q8 | 37 tok/s 22 tok/s | ~3.2s ~3.2s | |
#191 | MiniCPM5-2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
#192 | Qwen3.5-2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
#193 | Gemma 2 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
#194 | Qwen2.5-14B | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
#195 | OLMo 3.1 32B Think | 32B | INT4 Q8 | 28 tok/s 17 tok/s | ~3.3s ~3.6s | |
#196 | Nemotron 3.5 Lightning | 30B (3B active) | INT4 Q8 | 71 tok/s 66 tok/s | ~452 ms ~453 ms | |
#198 | Llama 3 70B | 70B | INT4 | 17 tok/s | ~7.2s | |
#200 | Gemma 4 E2B | 5.1B | INT4 Q8 FP16 | 131 tok/s 93 tok/s 56 tok/s | ~558 ms ~561 ms ~568 ms | |
#202 | DeepSeek-R1 8B | 8B | INT4 Q8 FP16 | 104 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#203 | Kimi Linear 48B A3B Instruct | 48B (3B active) | INT4 | 62 tok/s | ~488 ms | |
#204 | Qwen3-4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
#205 | Ministral 3 3B | 3B | INT4 Q8 FP16 | 126 tok/s 104 tok/s 73 tok/s | ~344 ms ~346 ms ~350 ms | |
#206 | OLMo 3 32B Base | 32B | INT4 Q8 | 28 tok/s 17 tok/s | ~3.3s ~3.6s | |
#207 | Gemma 2 9B | 9B | INT4 Q8 FP16 | 94 tok/s 62 tok/s 35 tok/s | ~961 ms ~966 ms ~979 ms | |
#208 | Qwen3.5-0.8B | 0.8B | INT4 Q8 FP16 | 238 tok/s 213 tok/s 173 tok/s | ~103 ms ~103 ms ~104 ms | |
#209 | Qwen3-14B | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
#210 | Gemma 3 4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
#211 | Qwen3-8B | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#213 | DeepSeek-R1 7B | 7B | INT4 Q8 FP16 | 112 tok/s 76 tok/s 44 tok/s | ~753 ms ~758 ms ~767 ms | |
#214 | Qwen3-1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 166 tok/s 120 tok/s | ~206 ms ~207 ms ~210 ms | |
#215 | Llama 3.1 8B | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#216 | Gemma 3 270M | 0.27B | INT4 Q8 FP16 | 256 tok/s 247 tok/s 226 tok/s | ~40 ms ~40 ms ~40 ms | |
#217 | Command R | 35B | INT4 Q8 | 26 tok/s 16 tok/s | ~3.6s ~4.0s | |
#218 | Qwen3-0.6B | 0.6B | INT4 Q8 FP16 | 249 tok/s 228 tok/s 192 tok/s | ~78 ms ~78 ms ~79 ms | |
#219 | Llama 3 8B | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#221 | Phi-4-Mini | 3.8B | INT4 Q8 FP16 | 151 tok/s 112 tok/s 71 tok/s | ~424 ms ~427 ms ~432 ms | |
#222 | Mixtral-8x7B-v0.1 | 46.7B (7B active) | INT4 Q8 | 100 tok/s 63 tok/s | ~873 ms ~1.0s | |
#223 | Qwen2.5-7B | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
#224 | Ministral-8B-2410 | 8B | INT4 Q8 FP16 | 102 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
#225 | DeepSeek-R1 1.5B | 1.5B | INT4 Q8 FP16 | 208 tok/s 175 tok/s 129 tok/s | ~184 ms ~185 ms ~187 ms | |
#226 | Llama 3.2 3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
#227 | Nemotron 3.5 Lightning 30B A3B | 30B (3B active) | INT4 Q8 | 148 tok/s 117 tok/s | ~444 ms ~446 ms | |
#228 | Phi-3-mini | 3.8B | INT4 Q8 FP16 | 151 tok/s 112 tok/s 71 tok/s | ~424 ms ~427 ms ~432 ms | |
#229 | Gemma 3n E2B IT | 6B (2B active) | INT4 Q8 FP16 | 126 tok/s 112 tok/s 86 tok/s | ~258 ms ~259 ms ~262 ms | |
#230 | Mistral-7B-Instruct-v0.1 | 7.3B | INT4 Q8 FP16 | 107 tok/s 72 tok/s 42 tok/s | ~784 ms ~788 ms ~798 ms | |
#231 | Llama 3.2 1B | 1B | INT4 Q8 FP16 | 228 tok/s 200 tok/s 157 tok/s | ~128 ms ~129 ms ~130 ms | |
#232 | Gemma 2 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
#233 | Phi-3-medium | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
#234 | OLMo 3 7B Instruct | 7B | INT4 Q8 FP16 | 84 tok/s 63 tok/s 39 tok/s | ~756 ms ~761 ms ~770 ms | |
#235 | Yi-34B | 34B | INT4 Q8 | 27 tok/s 16 tok/s | ~3.5s ~3.8s | |
#236 | Qwen2-7B | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
#237 | Phi-3-small | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
#238 | Gemma 3 1B | 1B | INT4 Q8 FP16 | 228 tok/s 200 tok/s 157 tok/s | ~128 ms ~129 ms ~130 ms | |
#239 | Mistral-7B-Instruct-v0.2 | 7.3B | INT4 Q8 FP16 | 107 tok/s 72 tok/s 42 tok/s | ~784 ms ~788 ms ~798 ms | |
#241 | Gemma 1 7B | 7B | INT4 Q8 FP16 | 84 tok/s 63 tok/s 39 tok/s | ~756 ms ~761 ms ~770 ms | |
#242 | Gemma 1 2B | 2B | INT4 Q8 FP16 | 198 tok/s 159 tok/s 111 tok/s | ~240 ms ~241 ms ~244 ms | |
#243 | ChatGLM3-6B | 6B | INT4 Q8 FP16 | 92 tok/s 70 tok/s 44 tok/s | ~653 ms ~657 ms ~665 ms | |
#244 | ChatGLM2-6B | 6B | INT4 Q8 FP16 | 92 tok/s 70 tok/s 44 tok/s | ~653 ms ~657 ms ~665 ms | |
#245 | ChatGLM-6B | 6B | INT4 Q8 FP16 | 92 tok/s 70 tok/s 44 tok/s | ~653 ms ~657 ms ~665 ms | |
Unranked Models (Sorted by Speed) | ||||||
- | DINOv3 ViT-L/16 | 0.3B | INT4 Q8 FP16 | 267 tok/s 254 tok/s 230 tok/s | ~43 ms ~43 ms ~44 ms | |
- | Qwen2-0.5B | 0.5B | INT4 Q8 FP16 | 255 tok/s 236 tok/s 204 tok/s | ~66 ms ~66 ms ~67 ms | |
- | Qwen2.5-0.5B | 0.5B | INT4 Q8 FP16 | 255 tok/s 236 tok/s 204 tok/s | ~66 ms ~66 ms ~67 ms | |
- | ERNIE-4.5-0.3B | 0.3B | INT4 Q8 FP16 | 253 tok/s 242 tok/s 221 tok/s | ~43 ms ~43 ms ~44 ms | |
- | ERNIE-4.5-0.3B-Base | 0.3B | INT4 Q8 FP16 | 253 tok/s 242 tok/s 221 tok/s | ~43 ms ~43 ms ~44 ms | |
- | Liquid D1-Omni 600M | 0.6B | INT4 Q8 FP16 | 249 tok/s 228 tok/s 192 tok/s | ~78 ms ~78 ms ~79 ms | |
- | PixArt-α XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 249 tok/s 228 tok/s 191 tok/s | ~79 ms ~79 ms ~80 ms | |
- | PixArt-Σ XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 249 tok/s 228 tok/s 191 tok/s | ~79 ms ~79 ms ~80 ms | |
- | Nemotron 3.5 Parse | 0.63B | INT4 Q8 FP16 | 247 tok/s 226 tok/s 189 tok/s | ~82 ms ~82 ms ~83 ms | |
- | SD 2.1 Turbo | 0.86B | INT4 Q8 FP16 | 235 tok/s 209 tok/s 168 tok/s | ~110 ms ~110 ms ~112 ms | |
- | Stable Diffusion 1.4 | 0.86B | INT4 Q8 FP16 | 235 tok/s 209 tok/s 168 tok/s | ~110 ms ~110 ms ~112 ms | |
- | Stable Diffusion 2.1 | 0.86B | INT4 Q8 FP16 | 235 tok/s 209 tok/s 168 tok/s | ~110 ms ~110 ms ~112 ms | |
- | Falcon-1B | 1B | INT4 Q8 FP16 | 234 tok/s 204 tok/s 160 tok/s | ~128 ms ~129 ms ~130 ms | |
- | Falcon3-1B | 1B | INT4 Q8 FP16 | 228 tok/s 200 tok/s 157 tok/s | ~128 ms ~129 ms ~130 ms | |
- | MiniCPM-V 4.7 1B | 1B | INT4 Q8 FP16 | 228 tok/s 200 tok/s 157 tok/s | ~128 ms ~129 ms ~130 ms | |
- | Kandinsky 2.2 Decoder | 1.22B | INT4 Q8 FP16 | 219 tok/s 188 tok/s 143 tok/s | ~153 ms ~153 ms ~155 ms | |
- | Wan 2.1 T2V 1.3B | 1.3B | INT4 Q8 FP16 | 216 tok/s 184 tok/s 139 tok/s | ~161 ms ~162 ms ~164 ms | |
- | Qwen2-1.5B | 1.5B | INT4 Q8 FP16 | 208 tok/s 175 tok/s 129 tok/s | ~184 ms ~185 ms ~187 ms | |
- | Qwen2.5-1.5B | 1.5B | INT4 Q8 FP16 | 208 tok/s 175 tok/s 129 tok/s | ~184 ms ~185 ms ~187 ms | |
- | SVD Image-to-Video XT | 1.5B | INT4 Q8 FP16 | 208 tok/s 175 tok/s 129 tok/s | ~184 ms ~185 ms ~187 ms | |
- | SVD Image-to-Video XT 1.1 | 1.5B | INT4 Q8 FP16 | 208 tok/s 175 tok/s 129 tok/s | ~184 ms ~185 ms ~187 ms | |
- | LFM3 1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 166 tok/s 120 tok/s | ~206 ms ~207 ms ~210 ms | |
- | LFM3 1.7B Base | 1.7B | INT4 Q8 FP16 | 201 tok/s 166 tok/s 120 tok/s | ~206 ms ~207 ms ~210 ms | |
- | Spark X2.5 1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 166 tok/s 120 tok/s | ~206 ms ~207 ms ~210 ms | |
- | Hy-MT2 1.8B | 1.8B | INT4 Q8 FP16 | 198 tok/s 162 tok/s 116 tok/s | ~217 ms ~218 ms ~220 ms | |
- | CogVideoX 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | LFM3 Audio 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | LFM3 Omni 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | LFM3 VL 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | Liquid D1 2B | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | LTX-Video 0.9.1 | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | Moondream 3.1 | 9B (2B active) | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~269 ms ~270 ms ~273 ms | |
- | Stable Diffusion 3 Medium | 2B | INT4 Q8 FP16 | 191 tok/s 155 tok/s 109 tok/s | ~240 ms ~241 ms ~244 ms | |
- | Kandinsky 3.0 | 2.1B | INT4 Q8 FP16 | 189 tok/s 152 tok/s 106 tok/s | ~250 ms ~252 ms ~254 ms | |
- | LTX-2 | 2.2B | INT4 Q8 FP16 | 186 tok/s 149 tok/s 103 tok/s | ~260 ms ~262 ms ~265 ms | |
- | SDXL 1.0 Refiner | 2.31B | INT4 Q8 FP16 | 183 tok/s 145 tok/s 100 tok/s | ~272 ms ~273 ms ~276 ms | |
- | Stable Diffusion 3.5 Medium | 2.5B | INT4 Q8 FP16 | 178 tok/s 140 tok/s 95 tok/s | ~291 ms ~293 ms ~296 ms | |
- | SDXL 1.0 Base | 2.56B | INT4 Q8 FP16 | 177 tok/s 138 tok/s 93 tok/s | ~297 ms ~299 ms ~302 ms | |
- | SDXL Turbo | 2.56B | INT4 Q8 FP16 | 177 tok/s 138 tok/s 93 tok/s | ~297 ms ~299 ms ~302 ms | |
- | Falcon-3B | 3B | INT4 Q8 FP16 | 173 tok/s 131 tok/s 85 tok/s | ~342 ms ~344 ms ~348 ms | |
- | CroissantLLM Base | 1.3B | INT4 Q8 FP16 | 171 tok/s 153 tok/s 120 tok/s | ~162 ms ~163 ms ~165 ms | |
- | Phi-1 | 1.3B | INT4 Q8 FP16 | 171 tok/s 153 tok/s 120 tok/s | ~162 ms ~163 ms ~165 ms | |
- | Phi-1.5 | 1.3B | INT4 Q8 FP16 | 171 tok/s 153 tok/s 120 tok/s | ~162 ms ~163 ms ~165 ms | |
- | DeepSeek-R1 3B | 3B | INT4 Q8 FP16 | 168 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | Mellum 2.1 12B A2.5B Thinking | 12B (2.5B active) | INT4 Q8 FP16 | 168 tok/s 134 tok/s 92 tok/s | ~334 ms ~336 ms ~339 ms | |
- | Falcon3-3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | LFM2-VL 3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | Liquid D1 3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | Liquid D1-VL 3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | Liquid LFM2-VL 3B D1 | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | Qwen2.5-3B | 3B | INT4 Q8 FP16 | 167 tok/s 128 tok/s 84 tok/s | ~342 ms ~344 ms ~348 ms | |
- | LFM2.5 24B A2B | 24B (2.5B active) | INT4 Q8 FP16 | 160 tok/s 129 tok/s 81 tok/s | ~378 ms ~379 ms ~448 ms | |
- | ERNIE-4.5-21B-A3B | 21B (3B active) | INT4 Q8 FP16 | 153 tok/s 120 tok/s 76 tok/s | ~416 ms ~418 ms ~458 ms | |
- | ERNIE-4.5-21B-A3B-Base | 21B (3B active) | INT4 Q8 FP16 | 153 tok/s 120 tok/s 76 tok/s | ~416 ms ~418 ms ~458 ms | |
- | ERNIE-4.5-VL-28B-A3B | 28B (3B active) | INT4 Q8 | 149 tok/s 118 tok/s | ~437 ms ~439 ms | |
- | ERNIE-4.5-VL-28B-A3B-Base | 28B (3B active) | INT4 Q8 | 149 tok/s 118 tok/s | ~437 ms ~439 ms | |
- | FLUX.2 Klein 4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
- | FrogNano 4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
- | Hy-Memory 30B A3B | 30B (3B active) | INT4 Q8 | 148 tok/s 117 tok/s | ~444 ms ~446 ms | |
- | Hy-MT2 30B A3B | 30B (3B active) | INT4 Q8 | 148 tok/s 117 tok/s | ~444 ms ~446 ms | |
- | Mage-VL Cowork | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
- | Spark X2.5 4B | 4B | INT4 Q8 FP16 | 148 tok/s 108 tok/s 68 tok/s | ~445 ms ~447 ms ~453 ms | |
- | MiniCPM-V 4.7 35B A3B | 35B (3B active) | INT4 Q8 | 146 tok/s 116 tok/s | ~453 ms ~455 ms | |
- | EmbeddingGemma 2 | 2B | INT4 Q8 FP16 | 144 tok/s 124 tok/s 93 tok/s | ~242 ms ~243 ms ~246 ms | |
- | CogVideoX 1.5 5B | 5B | INT4 Q8 FP16 | 133 tok/s 94 tok/s 57 tok/s | ~548 ms ~551 ms ~558 ms | |
- | CogVideoX 5B | 5B | INT4 Q8 FP16 | 133 tok/s 94 tok/s 57 tok/s | ~548 ms ~551 ms ~558 ms | |
- | Wan 2.2 TI2V 5B | 5B | INT4 Q8 FP16 | 133 tok/s 94 tok/s 57 tok/s | ~548 ms ~551 ms ~558 ms | |
- | Phi-2 | 2.7B | INT4 Q8 FP16 | 131 tok/s 110 tok/s 78 tok/s | ~313 ms ~315 ms ~318 ms | |
- | MaLLaM-3B | 3B | INT4 Q8 FP16 | 126 tok/s 104 tok/s 73 tok/s | ~344 ms ~346 ms ~350 ms | |
- | SmolLM3 3B | 3B | INT4 Q8 FP16 | 126 tok/s 104 tok/s 73 tok/s | ~344 ms ~346 ms ~350 ms | |
- | AliceAI Foundation 80B A3B Base | 80B (3B active) | INT4 | 119 tok/s | ~659 ms | |
- | Falcon-7B | 7B | INT4 Q8 FP16 | 114 tok/s 76 tok/s 44 tok/s | ~753 ms ~758 ms ~767 ms | |
- | Kolibri 1 | 78B (3.5B active) | INT4 | 114 tok/s | ~706 ms | |
- | AuraFlow v0.2 | 6.8B | INT4 Q8 FP16 | 112 tok/s 76 tok/s 45 tok/s | ~733 ms ~737 ms ~747 ms | |
- | AuraFlow v0.3 | 6.8B | INT4 Q8 FP16 | 112 tok/s 76 tok/s 45 tok/s | ~733 ms ~737 ms ~747 ms | |
- | Falcon3-7B | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
- | Hy-MT2 7B | 7B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 44 tok/s | ~754 ms ~758 ms ~767 ms | |
- | Mistral-7B-v0.1 | 7.3B | INT4 Q8 FP16 | 107 tok/s 72 tok/s 42 tok/s | ~784 ms ~788 ms ~798 ms | |
- | Youtu Parsing Omni | 8B | INT4 Q8 FP16 | 104 tok/s 68 tok/s 39 tok/s | ~856 ms ~861 ms ~872 ms | |
- | Stable Diffusion 3.5 Large | 8.1B | INT4 Q8 FP16 | 101 tok/s 67 tok/s 39 tok/s | ~866 ms ~871 ms ~882 ms | |
- | Stable Diffusion 3.5 Large Turbo | 8.1B | INT4 Q8 FP16 | 101 tok/s 67 tok/s 39 tok/s | ~866 ms ~871 ms ~882 ms | |
- | FLUX.2 Klein 9B | 9B | INT4 Q8 FP16 | 94 tok/s 62 tok/s 35 tok/s | ~961 ms ~966 ms ~979 ms | |
- | LensVLM 9B | 9B | INT4 Q8 FP16 | 94 tok/s 62 tok/s 35 tok/s | ~961 ms ~966 ms ~979 ms | |
- | MiMo V2.6 Distill Qwen 9B | 9B | INT4 Q8 FP16 | 94 tok/s 62 tok/s 35 tok/s | ~961 ms ~966 ms ~979 ms | |
- | ChatGLM3-6B-32K | 6B | INT4 Q8 FP16 | 92 tok/s 70 tok/s 44 tok/s | ~653 ms ~657 ms ~665 ms | |
- | Yi-6B | 6B | INT4 Q8 FP16 | 92 tok/s 70 tok/s 44 tok/s | ~653 ms ~657 ms ~665 ms | |
- | Falcon3-10B | 10B | INT4 Q8 FP16 | 88 tok/s 57 tok/s 32 tok/s | ~1.1s ~1.1s ~1.1s | |
- | Kimi-VL-A3B-Instruct | 16B (3B active) | INT4 Q8 FP16 | 88 tok/s 78 tok/s 59 tok/s | ~405 ms ~406 ms ~410 ms | |
- | Mochi 1 Preview | 10B | INT4 Q8 FP16 | 88 tok/s 57 tok/s 32 tok/s | ~1.1s ~1.1s ~1.1s | |
- | Falcon2-11B | 11B | INT4 Q8 FP16 | 85 tok/s 54 tok/s 30 tok/s | ~1.2s ~1.2s ~1.2s | |
- | Hunyuan Lite | 7B | INT4 Q8 FP16 | 84 tok/s 63 tok/s 39 tok/s | ~756 ms ~761 ms ~770 ms | |
- | OLMo 3 7B Base | 7B | INT4 Q8 FP16 | 84 tok/s 63 tok/s 39 tok/s | ~756 ms ~761 ms ~770 ms | |
- | SEA-LION-7B | 7.1B | INT4 Q8 FP16 | 84 tok/s 62 tok/s 39 tok/s | ~766 ms ~771 ms ~780 ms | |
- | SEA-LION-7B-Instruct | 7.1B | INT4 Q8 FP16 | 84 tok/s 62 tok/s 39 tok/s | ~766 ms ~771 ms ~780 ms | |
- | FLUX.1 Canny Dev | 11.9B | INT4 Q8 FP16 | 78 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
- | FLUX.1 Depth Dev | 11.9B | INT4 Q8 FP16 | 78 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
- | FLUX.1 Dev | 11.9B | INT4 Q8 FP16 | 78 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
- | FLUX.1 Fill Dev | 11.9B | INT4 Q8 FP16 | 78 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
- | FLUX.1 Schnell | 11.9B | INT4 Q8 FP16 | 78 tok/s 49 tok/s 27 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Sahabat-AI-Llama3-8B-Instruct | 8B | INT4 Q8 FP16 | 78 tok/s 57 tok/s 35 tok/s | ~859 ms ~864 ms ~875 ms | |
- | Typhoon-2-8B | 8B | INT4 Q8 FP16 | 78 tok/s 57 tok/s 35 tok/s | ~859 ms ~864 ms ~875 ms | |
- | HunyuanVideo 1.5 720P T2V | 13B | INT4 Q8 FP16 | 73 tok/s 46 tok/s 25 tok/s | ~1.4s ~1.4s ~1.4s | |
- | GLM-4-9B | 9B | INT4 Q8 FP16 | 72 tok/s 52 tok/s 32 tok/s | ~964 ms ~969 ms ~982 ms | |
- | GLM-4-9B-Chat | 9B | INT4 Q8 FP16 | 72 tok/s 52 tok/s 32 tok/s | ~964 ms ~969 ms ~982 ms | |
- | GLM-4-9B-Chat-1M | 9B | INT4 Q8 FP16 | 72 tok/s 52 tok/s 32 tok/s | ~964 ms ~969 ms ~982 ms | |
- | Yi-9B | 9B | INT4 Q8 FP16 | 72 tok/s 52 tok/s 32 tok/s | ~964 ms ~969 ms ~982 ms | |
- | Sahabat-AI-Gemma2-9B | 9.2B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 31 tok/s | ~984 ms ~990 ms ~1.0s | |
- | Sahabat-AI-Gemma2-9B-Instruct | 9.2B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 31 tok/s | ~984 ms ~990 ms ~1.0s | |
- | Wan 2.1 I2V 14B 720P | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Wan 2.1 T2V 14B | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Wan 2.2 T2V A14B | 14B | INT4 Q8 FP16 | 69 tok/s 43 tok/s 24 tok/s | ~1.5s ~1.5s ~1.5s | |
- | FLUX.2 Dev | 15B | INT4 Q8 FP16 | 65 tok/s 41 tok/s 22 tok/s | ~1.6s ~1.6s ~1.6s | |
- | GreenMind-14B-R1 | 14B | INT4 Q8 FP16 | 52 tok/s 37 tok/s 22 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Arete 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
- | AREX-2 | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
- | PPLX Decider v1 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
- | PPLX Decider v1.1 27B | 27B | INT4 Q8 | 41 tok/s 24 tok/s | ~2.8s ~2.8s | |
- | Agnes 3.0 Flash | 33B | INT4 Q8 | 34 tok/s 20 tok/s | ~3.4s ~3.4s | |
- | Falcon-40B | 40B | INT4 Q8 | 30 tok/s 16 tok/s | ~4.1s ~4.5s | |
- | Typhoon-2-70B | 70B | INT4 | 13 tok/s | ~8.5s | |
- | Kimi-Dev-72B | 72B | INT4 | 12 tok/s | ~8.7s | |
Inference Capacity
Maximum model parameter class runnable by weight precision.
Context:
4-bit
32B Models
8-bit
32B 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
7B-8B Models
Workload Recommendations
Guidance for optimal precision and operational ceilings on Radeon PRO W7900 (48GB).
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 864 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 192 tok/s on an 8B Q4 model and 23 tok/s on a 70B Q4 model.
Frequently Asked Questions
©2025 ApX Machine Learning