Arc A770 (16GB)
The Arc A770 (16GB) provides 16 GB of memory with 560 GB/s bandwidth, supporting local inference via ROCm and Vulkan acceleration.
Memory
16 GB
Usable Memory Ceiling
16 GB
Bandwidth
560 GB/s
Max Inference Class
14B Models (Q4)
TDP
225W
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 Arc A770 (16GB)
Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on Arc A770 (16GB).
Sort by:
Context:
KV Cache:
GPUs:
| Rank | Model | Parameters | Precision | TPS (Approx.) | TTFT | Can Run |
|---|---|---|---|---|---|---|
#48 | Qwen 3.8 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
#93 | Qwen3.5-27B | 27B | INT4 | 26 tok/s | ~8.7s | |
#111 | MiMo V2 Flash | 15B (309B active) | INT4 | 3 tok/s | ~88.3s | |
#112 | Phi-4 Reasoning Plus | 14B | INT4 | 37 tok/s | ~3.9s | |
#117 | Gemma 4 26B A4B | 25.2B (3.8B active) | INT4 | 112 tok/s | ~1.4s | |
#125 | Qwen3.5-9B | 9B | INT4 Q8 | 73 tok/s 45 tok/s | ~2.5s ~2.5s | |
#128 | K2 Horizon 7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
#135 | Gemma 4 12B | 11.95B | INT4 Q8 | 42 tok/s 26 tok/s | ~3.4s ~4.0s | |
#140 | Qwen3.5-4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
#143 | GPT-OSS 20B | 21B (3.6B active) | INT4 | 53 tok/s | ~1.5s | |
#144 | GLM-4V | 9B | INT4 Q8 | 53 tok/s 35 tok/s | ~2.5s ~2.8s | |
#150 | NVIDIA Nemotron 3 Nano 30B-A3B | 3.5B (30B active) | INT4 Q8 FP16 | 25 tok/s 14 tok/s 8 tok/s | ~8.0s ~8.0s ~8.1s | |
#158 | Granite 4.2 8B | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#159 | Gemma 4 E4B | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#166 | Phi-4 | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
#167 | Gemma 3 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
#178 | OLMo 3 7B Think | 7B | INT4 Q8 | 64 tok/s 46 tok/s | ~2.0s ~2.0s | |
#179 | DeepSeek-R1 14B | 14B | INT4 Q8 | 52 tok/s 28 tok/s | ~3.9s ~4.6s | |
#183 | Ministral 3 14B | 14B | INT4 | 37 tok/s | ~3.9s | |
#186 | Gemma 3 12B | 12B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.4s ~3.7s | |
#188 | Mistral-Small-2501 | 24B | INT4 | 30 tok/s | ~7.2s | |
#189 | Ministral 3 8B | 8B | INT4 Q8 | 58 tok/s 41 tok/s | ~2.3s ~2.3s | |
#191 | MiniCPM5-2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
#192 | Qwen3.5-2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
#193 | Gemma 2 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
#194 | Qwen2.5-14B | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
#200 | Gemma 4 E2B | 5.1B | INT4 Q8 FP16 | 111 tok/s 72 tok/s 41 tok/s | ~1.5s ~1.5s ~1.5s | |
#202 | DeepSeek-R1 8B | 8B | INT4 Q8 | 82 tok/s 50 tok/s | ~2.3s ~2.3s | |
#204 | Qwen3-4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
#205 | Ministral 3 3B | 3B | INT4 Q8 FP16 | 105 tok/s 83 tok/s 54 tok/s | ~893 ms ~896 ms ~902 ms | |
#207 | Gemma 2 9B | 9B | INT4 Q8 | 73 tok/s 45 tok/s | ~2.5s ~2.5s | |
#208 | Qwen3.5-0.8B | 0.8B | INT4 Q8 FP16 | 265 tok/s 221 tok/s 161 tok/s | ~255 ms ~255 ms ~257 ms | |
#209 | Qwen3-14B | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
#210 | Gemma 3 4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
#211 | Qwen3-8B | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#213 | DeepSeek-R1 7B | 7B | INT4 Q8 FP16 | 91 tok/s 56 tok/s 29 tok/s | ~2.0s ~2.0s ~2.2s | |
#214 | Qwen3-1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 152 tok/s 98 tok/s | ~531 ms ~532 ms ~536 ms | |
#215 | Llama 3.1 8B | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#216 | Gemma 3 270M | 0.27B | INT4 Q8 FP16 | 301 tok/s 281 tok/s 243 tok/s | ~88 ms ~88 ms ~88 ms | |
#218 | Qwen3-0.6B | 0.6B | INT4 Q8 FP16 | 286 tok/s 246 tok/s 188 tok/s | ~190 ms ~190 ms ~191 ms | |
#219 | Llama 3 8B | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#221 | Phi-4-Mini | 3.8B | INT4 Q8 FP16 | 133 tok/s 90 tok/s 52 tok/s | ~1.1s ~1.1s ~1.1s | |
#223 | Qwen2.5-7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
#224 | Ministral-8B-2410 | 8B | INT4 Q8 | 80 tok/s 50 tok/s | ~2.3s ~2.3s | |
#225 | DeepSeek-R1 1.5B | 1.5B | INT4 Q8 FP16 | 212 tok/s 163 tok/s 108 tok/s | ~470 ms ~471 ms ~475 ms | |
#226 | Llama 3.2 3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
#228 | Phi-3-mini | 3.8B | INT4 Q8 FP16 | 133 tok/s 90 tok/s 52 tok/s | ~1.1s ~1.1s ~1.1s | |
#229 | Gemma 3n E2B IT | 6B (2B active) | INT4 Q8 FP16 | 105 tok/s 90 tok/s 62 tok/s | ~665 ms ~667 ms ~728 ms | |
#230 | Mistral-7B-Instruct-v0.1 | 7.3B | INT4 Q8 FP16 | 86 tok/s 54 tok/s 27 tok/s | ~2.1s ~2.1s ~2.4s | |
#231 | Llama 3.2 1B | 1B | INT4 Q8 FP16 | 246 tok/s 200 tok/s 141 tok/s | ~322 ms ~323 ms ~325 ms | |
#232 | Gemma 2 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
#233 | Phi-3-medium | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
#234 | OLMo 3 7B Instruct | 7B | INT4 Q8 | 64 tok/s 46 tok/s | ~2.0s ~2.0s | |
#236 | Qwen2-7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
#237 | Phi-3-small | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
#238 | Gemma 3 1B | 1B | INT4 Q8 FP16 | 246 tok/s 200 tok/s 141 tok/s | ~322 ms ~323 ms ~325 ms | |
#239 | Mistral-7B-Instruct-v0.2 | 7.3B | INT4 Q8 FP16 | 86 tok/s 54 tok/s 27 tok/s | ~2.1s ~2.1s ~2.4s | |
#241 | Gemma 1 7B | 7B | INT4 Q8 | 64 tok/s 46 tok/s | ~2.0s ~2.0s | |
#242 | Gemma 1 2B | 2B | INT4 Q8 FP16 | 197 tok/s 143 tok/s 89 tok/s | ~619 ms ~621 ms ~625 ms | |
#243 | ChatGLM3-6B | 6B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 30 tok/s | ~1.7s ~1.7s ~1.9s | |
#244 | ChatGLM2-6B | 6B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 30 tok/s | ~1.7s ~1.7s ~1.9s | |
#245 | ChatGLM-6B | 6B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 30 tok/s | ~1.7s ~1.7s ~1.9s | |
Unranked Models (Sorted by Speed) | ||||||
- | DINOv3 ViT-L/16 | 0.3B | INT4 Q8 FP16 | 324 tok/s 297 tok/s 250 tok/s | ~97 ms ~97 ms ~98 ms | |
- | Qwen2-0.5B | 0.5B | INT4 Q8 FP16 | 298 tok/s 261 tok/s 205 tok/s | ~158 ms ~158 ms ~159 ms | |
- | Qwen2.5-0.5B | 0.5B | INT4 Q8 FP16 | 298 tok/s 261 tok/s 205 tok/s | ~158 ms ~158 ms ~159 ms | |
- | ERNIE-4.5-0.3B | 0.3B | INT4 Q8 FP16 | 294 tok/s 273 tok/s 233 tok/s | ~97 ms ~97 ms ~98 ms | |
- | ERNIE-4.5-0.3B-Base | 0.3B | INT4 Q8 FP16 | 294 tok/s 273 tok/s 233 tok/s | ~97 ms ~97 ms ~98 ms | |
- | Liquid D1-Omni 600M | 0.6B | INT4 Q8 FP16 | 286 tok/s 246 tok/s 188 tok/s | ~190 ms ~190 ms ~191 ms | |
- | PixArt-α XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 285 tok/s 245 tok/s 186 tok/s | ~192 ms ~193 ms ~194 ms | |
- | PixArt-Σ XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 285 tok/s 245 tok/s 186 tok/s | ~192 ms ~193 ms ~194 ms | |
- | Nemotron 3.5 Parse | 0.63B | INT4 Q8 FP16 | 283 tok/s 242 tok/s 183 tok/s | ~200 ms ~200 ms ~202 ms | |
- | SD 2.1 Turbo | 0.86B | INT4 Q8 FP16 | 259 tok/s 214 tok/s 154 tok/s | ~274 ms ~275 ms ~277 ms | |
- | Stable Diffusion 1.4 | 0.86B | INT4 Q8 FP16 | 259 tok/s 214 tok/s 154 tok/s | ~274 ms ~275 ms ~277 ms | |
- | Stable Diffusion 2.1 | 0.86B | INT4 Q8 FP16 | 259 tok/s 214 tok/s 154 tok/s | ~274 ms ~275 ms ~277 ms | |
- | Falcon-1B | 1B | INT4 Q8 FP16 | 257 tok/s 206 tok/s 144 tok/s | ~322 ms ~323 ms ~325 ms | |
- | Falcon3-1B | 1B | INT4 Q8 FP16 | 246 tok/s 200 tok/s 141 tok/s | ~322 ms ~323 ms ~325 ms | |
- | MiniCPM-V 4.7 1B | 1B | INT4 Q8 FP16 | 246 tok/s 200 tok/s 141 tok/s | ~322 ms ~323 ms ~325 ms | |
- | Kandinsky 2.2 Decoder | 1.22B | INT4 Q8 FP16 | 230 tok/s 182 tok/s 124 tok/s | ~388 ms ~389 ms ~391 ms | |
- | Wan 2.1 T2V 1.3B | 1.3B | INT4 Q8 FP16 | 225 tok/s 176 tok/s 119 tok/s | ~410 ms ~411 ms ~414 ms | |
- | Qwen2-1.5B | 1.5B | INT4 Q8 FP16 | 212 tok/s 163 tok/s 108 tok/s | ~470 ms ~471 ms ~475 ms | |
- | Qwen2.5-1.5B | 1.5B | INT4 Q8 FP16 | 212 tok/s 163 tok/s 108 tok/s | ~470 ms ~471 ms ~475 ms | |
- | SVD Image-to-Video XT | 1.5B | INT4 Q8 FP16 | 212 tok/s 163 tok/s 108 tok/s | ~470 ms ~471 ms ~475 ms | |
- | SVD Image-to-Video XT 1.1 | 1.5B | INT4 Q8 FP16 | 212 tok/s 163 tok/s 108 tok/s | ~470 ms ~471 ms ~475 ms | |
- | LFM3 1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 152 tok/s 98 tok/s | ~531 ms ~532 ms ~536 ms | |
- | LFM3 1.7B Base | 1.7B | INT4 Q8 FP16 | 201 tok/s 152 tok/s 98 tok/s | ~531 ms ~532 ms ~536 ms | |
- | Spark X2.5 1.7B | 1.7B | INT4 Q8 FP16 | 201 tok/s 152 tok/s 98 tok/s | ~531 ms ~532 ms ~536 ms | |
- | Hy-MT2 1.8B | 1.8B | INT4 Q8 FP16 | 196 tok/s 147 tok/s 95 tok/s | ~558 ms ~560 ms ~564 ms | |
- | Moondream 3.1 | 9B (2B active) | INT4 Q8 | 187 tok/s 138 tok/s | ~698 ms ~700 ms | |
- | CogVideoX 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | LFM3 Audio 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | LFM3 Omni 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | LFM3 VL 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | Liquid D1 2B | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | LTX-Video 0.9.1 | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | Stable Diffusion 3 Medium | 2B | INT4 Q8 FP16 | 186 tok/s 138 tok/s 87 tok/s | ~620 ms ~622 ms ~626 ms | |
- | Kandinsky 3.0 | 2.1B | INT4 Q8 FP16 | 182 tok/s 134 tok/s 84 tok/s | ~647 ms ~649 ms ~653 ms | |
- | LTX-2 | 2.2B | INT4 Q8 FP16 | 178 tok/s 130 tok/s 81 tok/s | ~674 ms ~676 ms ~680 ms | |
- | SDXL 1.0 Refiner | 2.31B | INT4 Q8 FP16 | 174 tok/s 126 tok/s 78 tok/s | ~704 ms ~706 ms ~711 ms | |
- | Stable Diffusion 3.5 Medium | 2.5B | INT4 Q8 FP16 | 168 tok/s 120 tok/s 74 tok/s | ~755 ms ~757 ms ~763 ms | |
- | SDXL 1.0 Base | 2.56B | INT4 Q8 FP16 | 166 tok/s 118 tok/s 72 tok/s | ~771 ms ~774 ms ~779 ms | |
- | SDXL Turbo | 2.56B | INT4 Q8 FP16 | 166 tok/s 118 tok/s 72 tok/s | ~771 ms ~774 ms ~779 ms | |
- | Falcon-3B | 3B | INT4 Q8 FP16 | 161 tok/s 110 tok/s 65 tok/s | ~890 ms ~893 ms ~899 ms | |
- | CroissantLLM Base | 1.3B | INT4 Q8 FP16 | 158 tok/s 135 tok/s 99 tok/s | ~412 ms ~413 ms ~415 ms | |
- | Phi-1 | 1.3B | INT4 Q8 FP16 | 158 tok/s 135 tok/s 99 tok/s | ~412 ms ~413 ms ~415 ms | |
- | Phi-1.5 | 1.3B | INT4 Q8 FP16 | 158 tok/s 135 tok/s 99 tok/s | ~412 ms ~413 ms ~415 ms | |
- | DeepSeek-R1 3B | 3B | INT4 Q8 FP16 | 155 tok/s 107 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | Mellum 2.1 12B A2.5B Thinking | 12B (2.5B active) | INT4 Q8 | 155 tok/s 108 tok/s | ~870 ms ~947 ms | |
- | Falcon3-3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | LFM2-VL 3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | Liquid D1 3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | Liquid D1-VL 3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | Liquid LFM2-VL 3B D1 | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | Qwen2.5-3B | 3B | INT4 Q8 FP16 | 153 tok/s 106 tok/s 64 tok/s | ~890 ms ~893 ms ~900 ms | |
- | LFM2.5 24B A2B | 24B (2.5B active) | INT4 | 136 tok/s | ~1.1s | |
- | FLUX.2 Klein 4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
- | FrogNano 4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
- | Mage-VL Cowork | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
- | Spark X2.5 4B | 4B | INT4 Q8 FP16 | 129 tok/s 87 tok/s 50 tok/s | ~1.2s ~1.2s ~1.2s | |
- | ERNIE-4.5-21B-A3B | 21B (3B active) | INT4 | 128 tok/s | ~1.2s | |
- | ERNIE-4.5-21B-A3B-Base | 21B (3B active) | INT4 | 128 tok/s | ~1.2s | |
- | EmbeddingGemma 2 | 2B | INT4 Q8 FP16 | 124 tok/s 103 tok/s 72 tok/s | ~622 ms ~624 ms ~628 ms | |
- | ERNIE-4.5-VL-28B-A3B | 28B (3B active) | INT4 | 118 tok/s | ~1.3s | |
- | ERNIE-4.5-VL-28B-A3B-Base | 28B (3B active) | INT4 | 118 tok/s | ~1.3s | |
- | CogVideoX 1.5 5B | 5B | INT4 Q8 FP16 | 112 tok/s 73 tok/s 41 tok/s | ~1.4s ~1.4s ~1.5s | |
- | CogVideoX 5B | 5B | INT4 Q8 FP16 | 112 tok/s 73 tok/s 41 tok/s | ~1.4s ~1.4s ~1.5s | |
- | Wan 2.2 TI2V 5B | 5B | INT4 Q8 FP16 | 112 tok/s 73 tok/s 41 tok/s | ~1.4s ~1.4s ~1.5s | |
- | Phi-2 | 2.7B | INT4 Q8 FP16 | 110 tok/s 88 tok/s 59 tok/s | ~812 ms ~814 ms ~820 ms | |
- | MaLLaM-3B | 3B | INT4 Q8 FP16 | 105 tok/s 83 tok/s 54 tok/s | ~893 ms ~896 ms ~902 ms | |
- | SmolLM3 3B | 3B | INT4 Q8 FP16 | 105 tok/s 83 tok/s 54 tok/s | ~893 ms ~896 ms ~902 ms | |
- | Falcon-7B | 7B | INT4 Q8 FP16 | 92 tok/s 57 tok/s 29 tok/s | ~2.0s ~2.0s ~2.2s | |
- | AuraFlow v0.2 | 6.8B | INT4 Q8 FP16 | 90 tok/s 57 tok/s 30 tok/s | ~1.9s ~1.9s ~2.1s | |
- | AuraFlow v0.3 | 6.8B | INT4 Q8 FP16 | 90 tok/s 57 tok/s 30 tok/s | ~1.9s ~1.9s ~2.1s | |
- | Falcon3-7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
- | Hy-MT2 7B | 7B | INT4 Q8 FP16 | 88 tok/s 56 tok/s 28 tok/s | ~2.0s ~2.0s ~2.4s | |
- | Mistral-7B-v0.1 | 7.3B | INT4 Q8 FP16 | 86 tok/s 54 tok/s 27 tok/s | ~2.1s ~2.1s ~2.4s | |
- | Youtu Parsing Omni | 8B | INT4 Q8 | 82 tok/s 50 tok/s | ~2.3s ~2.3s | |
- | Stable Diffusion 3.5 Large | 8.1B | INT4 Q8 | 79 tok/s 49 tok/s | ~2.3s ~2.3s | |
- | Stable Diffusion 3.5 Large Turbo | 8.1B | INT4 Q8 | 79 tok/s 49 tok/s | ~2.3s ~2.3s | |
- | FLUX.2 Klein 9B | 9B | INT4 Q8 | 73 tok/s 45 tok/s | ~2.5s ~2.5s | |
- | LensVLM 9B | 9B | INT4 Q8 | 73 tok/s 45 tok/s | ~2.5s ~2.5s | |
- | MiMo V2.6 Distill Qwen 9B | 9B | INT4 Q8 | 73 tok/s 45 tok/s | ~2.5s ~2.5s | |
- | ChatGLM3-6B-32K | 6B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 30 tok/s | ~1.7s ~1.7s ~1.9s | |
- | Yi-6B | 6B | INT4 Q8 FP16 | 71 tok/s 51 tok/s 30 tok/s | ~1.7s ~1.7s ~1.9s | |
- | Falcon3-10B | 10B | INT4 Q8 | 67 tok/s 41 tok/s | ~2.8s ~2.8s | |
- | Mochi 1 Preview | 10B | INT4 Q8 | 67 tok/s 41 tok/s | ~2.8s ~2.8s | |
- | Falcon2-11B | 11B | INT4 Q8 | 65 tok/s 38 tok/s | ~3.1s ~3.1s | |
- | Hunyuan Lite | 7B | INT4 Q8 | 64 tok/s 46 tok/s | ~2.0s ~2.0s | |
- | Kimi-VL-A3B-Instruct | 16B (3B active) | INT4 | 64 tok/s | ~1.1s | |
- | OLMo 3 7B Base | 7B | INT4 Q8 | 64 tok/s 46 tok/s | ~2.0s ~2.0s | |
- | SEA-LION-7B | 7.1B | INT4 Q8 | 64 tok/s 45 tok/s | ~2.0s ~2.0s | |
- | SEA-LION-7B-Instruct | 7.1B | INT4 Q8 | 64 tok/s 45 tok/s | ~2.0s ~2.0s | |
- | FLUX.1 Canny Dev | 11.9B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.3s ~3.6s | |
- | FLUX.1 Depth Dev | 11.9B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.3s ~3.6s | |
- | FLUX.1 Dev | 11.9B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.3s ~3.6s | |
- | FLUX.1 Fill Dev | 11.9B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.3s ~3.6s | |
- | FLUX.1 Schnell | 11.9B | INT4 Q8 | 58 tok/s 33 tok/s | ~3.3s ~3.6s | |
- | Sahabat-AI-Llama3-8B-Instruct | 8B | INT4 Q8 | 58 tok/s 41 tok/s | ~2.3s ~2.3s | |
- | Typhoon-2-8B | 8B | INT4 Q8 | 58 tok/s 41 tok/s | ~2.3s ~2.3s | |
- | HunyuanVideo 1.5 720P T2V | 13B | INT4 Q8 | 54 tok/s 31 tok/s | ~3.6s ~4.0s | |
- | GLM-4-9B | 9B | INT4 Q8 | 53 tok/s 35 tok/s | ~2.5s ~2.8s | |
- | GLM-4-9B-Chat | 9B | INT4 Q8 | 53 tok/s 35 tok/s | ~2.5s ~2.8s | |
- | GLM-4-9B-Chat-1M | 9B | INT4 Q8 | 53 tok/s 35 tok/s | ~2.5s ~2.8s | |
- | Yi-9B | 9B | INT4 Q8 | 53 tok/s 35 tok/s | ~2.5s ~2.8s | |
- | Sahabat-AI-Gemma2-9B | 9.2B | INT4 Q8 | 52 tok/s 35 tok/s | ~2.6s ~2.8s | |
- | Sahabat-AI-Gemma2-9B-Instruct | 9.2B | INT4 Q8 | 52 tok/s 35 tok/s | ~2.6s ~2.8s | |
- | Wan 2.1 I2V 14B 720P | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
- | Wan 2.1 T2V 14B | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
- | Wan 2.2 T2V A14B | 14B | INT4 Q8 | 51 tok/s 27 tok/s | ~3.9s ~4.6s | |
- | FLUX.2 Dev | 15B | INT4 Q8 | 48 tok/s 26 tok/s | ~4.2s ~4.9s | |
- | GreenMind-14B-R1 | 14B | INT4 | 37 tok/s | ~3.9s | |
- | Arete 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
- | AREX-2 | 27B | INT4 | 26 tok/s | ~8.7s | |
- | PPLX Decider v1 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
- | PPLX Decider v1.1 27B | 27B | INT4 | 26 tok/s | ~8.7s | |
Inference Capacity
Maximum model parameter class runnable by weight precision.
Context:
4-bit
14B Models
8-bit
8B Models
16-bit
3B Models
Fine-Tuning Capacity
Hardware capacity for local training and adapter fine-tuning.
Context Length:
QLoRA
14B Models
LoRA
8B Models
Full Parameter Training
1B Models
Workload Recommendations
Guidance for optimal precision and operational ceilings on Arc A770 (16GB).
Inference Sweet Spot
Optimal for 8B models across extended context (32k+). 14B models fit comfortably at Q4 precision with standard context.
Bandwidth & Speed Profile
With 560 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 124 tok/s on an 8B Q4 model and 15 tok/s on a 70B Q4 model.
Frequently Asked Questions
©2025 ApX Machine Learning