AMD RX 6700 XT (12GB)
The AMD RX 6700 XT (12GB) provides 12 GB of memory with 384 GB/s bandwidth, supporting local inference via ROCm and Vulkan acceleration.
Memory
12 GB
Usable Memory Ceiling
12 GB
Bandwidth
384 GB/s
Max Inference Class
14B Models (Q4)
TDP
230W
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 AMD RX 6700 XT (12GB)
Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on AMD RX 6700 XT (12GB).
Sort by:
Context:
KV Cache:
GPUs:
| Rank | Model | Parameters | Precision | TPS (Approx.) | TTFT | Can Run |
|---|---|---|---|---|---|---|
#112 | Phi-4 Reasoning Plus | 14B | INT4 | 23 tok/s | ~7.9s | |
#125 | Qwen3.5-9B | 9B | INT4 Q8 | 51 tok/s 30 tok/s | ~4.4s ~4.8s | |
#128 | K2 Horizon 7B | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
#135 | Gemma 4 12B | 11.95B | INT4 | 28 tok/s | ~6.3s | |
#140 | Qwen3.5-4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
#144 | GLM-4V | 9B | INT4 | 37 tok/s | ~4.4s | |
#150 | NVIDIA Nemotron 3 Nano 30B-A3B | 3.5B (30B active) | INT4 Q8 FP16 | 17 tok/s 10 tok/s 5 tok/s | ~13.8s ~13.8s ~15.1s | |
#158 | Granite 4.2 8B | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#159 | Gemma 4 E4B | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#166 | Phi-4 | 14B | INT4 | 35 tok/s | ~6.7s | |
#178 | OLMo 3 7B Think | 7B | INT4 Q8 | 45 tok/s 30 tok/s | ~3.4s ~3.7s | |
#179 | DeepSeek-R1 14B | 14B | INT4 | 36 tok/s | ~6.7s | |
#183 | Ministral 3 14B | 14B | INT4 | 23 tok/s | ~7.9s | |
#186 | Gemma 3 12B | 12B | INT4 | 40 tok/s | ~5.8s | |
#189 | Ministral 3 8B | 8B | INT4 Q8 | 41 tok/s 26 tok/s | ~3.9s ~4.6s | |
#191 | MiniCPM5-2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
#192 | Qwen3.5-2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
#194 | Qwen2.5-14B | 14B | INT4 | 35 tok/s | ~6.7s | |
#200 | Gemma 4 E2B | 5.1B | INT4 Q8 FP16 | 78 tok/s 50 tok/s 25 tok/s | ~2.5s ~2.5s ~3.0s | |
#202 | DeepSeek-R1 8B | 8B | INT4 Q8 | 58 tok/s 35 tok/s | ~3.9s ~3.9s | |
#204 | Qwen3-4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
#205 | Ministral 3 3B | 3B | INT4 Q8 FP16 | 75 tok/s 58 tok/s 38 tok/s | ~1.5s ~1.5s ~1.6s | |
#207 | Gemma 2 9B | 9B | INT4 Q8 | 51 tok/s 30 tok/s | ~4.4s ~4.8s | |
#208 | Qwen3.5-0.8B | 0.8B | INT4 Q8 FP16 | 197 tok/s 162 tok/s 116 tok/s | ~436 ms ~437 ms ~439 ms | |
#209 | Qwen3-14B | 14B | INT4 | 35 tok/s | ~6.7s | |
#210 | Gemma 3 4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
#211 | Qwen3-8B | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#213 | DeepSeek-R1 7B | 7B | INT4 Q8 | 64 tok/s 39 tok/s | ~3.4s ~3.4s | |
#214 | Qwen3-1.7B | 1.7B | INT4 Q8 FP16 | 147 tok/s 109 tok/s 70 tok/s | ~912 ms ~914 ms ~920 ms | |
#215 | Llama 3.1 8B | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#216 | Gemma 3 270M | 0.27B | INT4 Q8 FP16 | 227 tok/s 211 tok/s 179 tok/s | ~147 ms ~148 ms ~149 ms | |
#218 | Qwen3-0.6B | 0.6B | INT4 Q8 FP16 | 215 tok/s 182 tok/s 137 tok/s | ~323 ms ~324 ms ~326 ms | |
#219 | Llama 3 8B | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#221 | Phi-4-Mini | 3.8B | INT4 Q8 FP16 | 95 tok/s 63 tok/s 36 tok/s | ~1.9s ~1.9s ~1.9s | |
#223 | Qwen2.5-7B | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
#224 | Ministral-8B-2410 | 8B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.2s | |
#225 | DeepSeek-R1 1.5B | 1.5B | INT4 Q8 FP16 | 155 tok/s 118 tok/s 76 tok/s | ~807 ms ~809 ms ~814 ms | |
#226 | Llama 3.2 3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
#228 | Phi-3-mini | 3.8B | INT4 Q8 FP16 | 95 tok/s 63 tok/s 36 tok/s | ~1.9s ~1.9s ~1.9s | |
#229 | Gemma 3n E2B IT | 6B (2B active) | INT4 Q8 | 74 tok/s 64 tok/s | ~1.1s ~1.1s | |
#230 | Mistral-7B-Instruct-v0.1 | 7.3B | INT4 Q8 | 60 tok/s 37 tok/s | ~3.6s ~3.6s | |
#231 | Llama 3.2 1B | 1B | INT4 Q8 FP16 | 182 tok/s 146 tok/s 101 tok/s | ~553 ms ~554 ms ~557 ms | |
#232 | Gemma 2 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
#233 | Phi-3-medium | 14B | INT4 | 35 tok/s | ~6.7s | |
#234 | OLMo 3 7B Instruct | 7B | INT4 Q8 | 45 tok/s 30 tok/s | ~3.4s ~3.7s | |
#236 | Qwen2-7B | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
#237 | Phi-3-small | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
#238 | Gemma 3 1B | 1B | INT4 Q8 FP16 | 182 tok/s 146 tok/s 101 tok/s | ~553 ms ~554 ms ~557 ms | |
#239 | Mistral-7B-Instruct-v0.2 | 7.3B | INT4 Q8 | 60 tok/s 37 tok/s | ~3.6s ~3.6s | |
#241 | Gemma 1 7B | 7B | INT4 Q8 | 45 tok/s 30 tok/s | ~3.4s ~3.7s | |
#242 | Gemma 1 2B | 2B | INT4 Q8 FP16 | 144 tok/s 103 tok/s 63 tok/s | ~1.1s ~1.1s ~1.1s | |
#243 | ChatGLM3-6B | 6B | INT4 Q8 | 50 tok/s 36 tok/s | ~3.0s ~3.0s | |
#244 | ChatGLM2-6B | 6B | INT4 Q8 | 50 tok/s 36 tok/s | ~3.0s ~3.0s | |
#245 | ChatGLM-6B | 6B | INT4 Q8 | 50 tok/s 36 tok/s | ~3.0s ~3.0s | |
Unranked Models (Sorted by Speed) | ||||||
- | DINOv3 ViT-L/16 | 0.3B | INT4 Q8 FP16 | 246 tok/s 223 tok/s 185 tok/s | ~163 ms ~164 ms ~165 ms | |
- | Qwen2-0.5B | 0.5B | INT4 Q8 FP16 | 224 tok/s 194 tok/s 150 tok/s | ~269 ms ~269 ms ~271 ms | |
- | Qwen2.5-0.5B | 0.5B | INT4 Q8 FP16 | 224 tok/s 194 tok/s 150 tok/s | ~269 ms ~269 ms ~271 ms | |
- | ERNIE-4.5-0.3B | 0.3B | INT4 Q8 FP16 | 221 tok/s 204 tok/s 172 tok/s | ~164 ms ~164 ms ~165 ms | |
- | ERNIE-4.5-0.3B-Base | 0.3B | INT4 Q8 FP16 | 221 tok/s 204 tok/s 172 tok/s | ~164 ms ~164 ms ~165 ms | |
- | Liquid D1-Omni 600M | 0.6B | INT4 Q8 FP16 | 215 tok/s 182 tok/s 137 tok/s | ~323 ms ~324 ms ~326 ms | |
- | PixArt-α XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 214 tok/s 181 tok/s 135 tok/s | ~328 ms ~329 ms ~331 ms | |
- | PixArt-Σ XL 2 1024-MS | 0.61B | INT4 Q8 FP16 | 214 tok/s 181 tok/s 135 tok/s | ~328 ms ~329 ms ~331 ms | |
- | Nemotron 3.5 Parse | 0.63B | INT4 Q8 FP16 | 212 tok/s 179 tok/s 133 tok/s | ~341 ms ~342 ms ~344 ms | |
- | SD 2.1 Turbo | 0.86B | INT4 Q8 FP16 | 193 tok/s 157 tok/s 111 tok/s | ~470 ms ~471 ms ~473 ms | |
- | Stable Diffusion 1.4 | 0.86B | INT4 Q8 FP16 | 193 tok/s 157 tok/s 111 tok/s | ~470 ms ~471 ms ~473 ms | |
- | Stable Diffusion 2.1 | 0.86B | INT4 Q8 FP16 | 193 tok/s 157 tok/s 111 tok/s | ~470 ms ~471 ms ~473 ms | |
- | Falcon-1B | 1B | INT4 Q8 FP16 | 191 tok/s 151 tok/s 103 tok/s | ~552 ms ~554 ms ~557 ms | |
- | Falcon3-1B | 1B | INT4 Q8 FP16 | 182 tok/s 146 tok/s 101 tok/s | ~553 ms ~554 ms ~557 ms | |
- | MiniCPM-V 4.7 1B | 1B | INT4 Q8 FP16 | 182 tok/s 146 tok/s 101 tok/s | ~553 ms ~554 ms ~557 ms | |
- | Kandinsky 2.2 Decoder | 1.22B | INT4 Q8 FP16 | 169 tok/s 132 tok/s 88 tok/s | ~665 ms ~667 ms ~671 ms | |
- | Wan 2.1 T2V 1.3B | 1.3B | INT4 Q8 FP16 | 165 tok/s 128 tok/s 85 tok/s | ~703 ms ~705 ms ~709 ms | |
- | Qwen2-1.5B | 1.5B | INT4 Q8 FP16 | 155 tok/s 118 tok/s 76 tok/s | ~807 ms ~809 ms ~814 ms | |
- | Qwen2.5-1.5B | 1.5B | INT4 Q8 FP16 | 155 tok/s 118 tok/s 76 tok/s | ~807 ms ~809 ms ~814 ms | |
- | SVD Image-to-Video XT | 1.5B | INT4 Q8 FP16 | 155 tok/s 118 tok/s 76 tok/s | ~807 ms ~809 ms ~814 ms | |
- | SVD Image-to-Video XT 1.1 | 1.5B | INT4 Q8 FP16 | 155 tok/s 118 tok/s 76 tok/s | ~807 ms ~809 ms ~814 ms | |
- | LFM3 1.7B | 1.7B | INT4 Q8 FP16 | 147 tok/s 109 tok/s 70 tok/s | ~912 ms ~914 ms ~920 ms | |
- | LFM3 1.7B Base | 1.7B | INT4 Q8 FP16 | 147 tok/s 109 tok/s 70 tok/s | ~912 ms ~914 ms ~920 ms | |
- | Spark X2.5 1.7B | 1.7B | INT4 Q8 FP16 | 147 tok/s 109 tok/s 70 tok/s | ~912 ms ~914 ms ~920 ms | |
- | Hy-MT2 1.8B | 1.8B | INT4 Q8 FP16 | 143 tok/s 106 tok/s 67 tok/s | ~960 ms ~963 ms ~968 ms | |
- | CogVideoX 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | LFM3 Audio 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | LFM3 Omni 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | LFM3 VL 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | Liquid D1 2B | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | LTX-Video 0.9.1 | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | Moondream 3.1 | 9B (2B active) | INT4 Q8 | 135 tok/s 94 tok/s | ~1.2s ~1.3s | |
- | Stable Diffusion 3 Medium | 2B | INT4 Q8 FP16 | 135 tok/s 99 tok/s 61 tok/s | ~1.1s ~1.1s ~1.1s | |
- | Kandinsky 3.0 | 2.1B | INT4 Q8 FP16 | 132 tok/s 96 tok/s 59 tok/s | ~1.1s ~1.1s ~1.1s | |
- | LTX-2 | 2.2B | INT4 Q8 FP16 | 129 tok/s 93 tok/s 57 tok/s | ~1.2s ~1.2s ~1.2s | |
- | SDXL 1.0 Refiner | 2.31B | INT4 Q8 FP16 | 126 tok/s 90 tok/s 55 tok/s | ~1.2s ~1.2s ~1.2s | |
- | Stable Diffusion 3.5 Medium | 2.5B | INT4 Q8 FP16 | 121 tok/s 85 tok/s 52 tok/s | ~1.3s ~1.3s ~1.3s | |
- | SDXL 1.0 Base | 2.56B | INT4 Q8 FP16 | 120 tok/s 84 tok/s 51 tok/s | ~1.3s ~1.3s ~1.3s | |
- | SDXL Turbo | 2.56B | INT4 Q8 FP16 | 120 tok/s 84 tok/s 51 tok/s | ~1.3s ~1.3s ~1.3s | |
- | Falcon-3B | 3B | INT4 Q8 FP16 | 116 tok/s 78 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | CroissantLLM Base | 1.3B | INT4 Q8 FP16 | 114 tok/s 97 tok/s 70 tok/s | ~706 ms ~708 ms ~712 ms | |
- | Phi-1 | 1.3B | INT4 Q8 FP16 | 114 tok/s 97 tok/s 70 tok/s | ~706 ms ~708 ms ~712 ms | |
- | Phi-1.5 | 1.3B | INT4 Q8 FP16 | 114 tok/s 97 tok/s 70 tok/s | ~706 ms ~708 ms ~712 ms | |
- | DeepSeek-R1 3B | 3B | INT4 Q8 FP16 | 111 tok/s 76 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Mellum 2.1 12B A2.5B Thinking | 12B (2.5B active) | INT4 | 111 tok/s | ~1.5s | |
- | Falcon3-3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | LFM2-VL 3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Liquid D1 3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Liquid D1-VL 3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Liquid LFM2-VL 3B D1 | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | Qwen2.5-3B | 3B | INT4 Q8 FP16 | 110 tok/s 75 tok/s 45 tok/s | ~1.5s ~1.5s ~1.5s | |
- | FLUX.2 Klein 4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
- | FrogNano 4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
- | Mage-VL Cowork | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
- | Spark X2.5 4B | 4B | INT4 Q8 FP16 | 92 tok/s 61 tok/s 35 tok/s | ~2.0s ~2.0s ~2.0s | |
- | EmbeddingGemma 2 | 2B | INT4 Q8 FP16 | 88 tok/s 73 tok/s 50 tok/s | ~1.1s ~1.1s ~1.1s | |
- | CogVideoX 1.5 5B | 5B | INT4 Q8 FP16 | 79 tok/s 51 tok/s 27 tok/s | ~2.5s ~2.5s ~2.7s | |
- | CogVideoX 5B | 5B | INT4 Q8 FP16 | 79 tok/s 51 tok/s 27 tok/s | ~2.5s ~2.5s ~2.7s | |
- | Wan 2.2 TI2V 5B | 5B | INT4 Q8 FP16 | 79 tok/s 51 tok/s 27 tok/s | ~2.5s ~2.5s ~2.7s | |
- | Phi-2 | 2.7B | INT4 Q8 FP16 | 78 tok/s 62 tok/s 41 tok/s | ~1.4s ~1.4s ~1.4s | |
- | MaLLaM-3B | 3B | INT4 Q8 FP16 | 75 tok/s 58 tok/s 38 tok/s | ~1.5s ~1.5s ~1.6s | |
- | SmolLM3 3B | 3B | INT4 Q8 FP16 | 75 tok/s 58 tok/s 38 tok/s | ~1.5s ~1.5s ~1.6s | |
- | Falcon-7B | 7B | INT4 Q8 | 65 tok/s 40 tok/s | ~3.4s ~3.4s | |
- | AuraFlow v0.2 | 6.8B | INT4 Q8 | 64 tok/s 40 tok/s | ~3.3s ~3.3s | |
- | AuraFlow v0.3 | 6.8B | INT4 Q8 | 64 tok/s 40 tok/s | ~3.3s ~3.3s | |
- | Falcon3-7B | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
- | Hy-MT2 7B | 7B | INT4 Q8 | 62 tok/s 39 tok/s | ~3.4s ~3.4s | |
- | Mistral-7B-v0.1 | 7.3B | INT4 Q8 | 60 tok/s 37 tok/s | ~3.6s ~3.6s | |
- | Youtu Parsing Omni | 8B | INT4 Q8 | 58 tok/s 35 tok/s | ~3.9s ~3.9s | |
- | Stable Diffusion 3.5 Large | 8.1B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.3s | |
- | Stable Diffusion 3.5 Large Turbo | 8.1B | INT4 Q8 | 56 tok/s 33 tok/s | ~3.9s ~4.3s | |
- | FLUX.2 Klein 9B | 9B | INT4 Q8 | 51 tok/s 30 tok/s | ~4.4s ~4.8s | |
- | LensVLM 9B | 9B | INT4 Q8 | 51 tok/s 30 tok/s | ~4.4s ~4.8s | |
- | MiMo V2.6 Distill Qwen 9B | 9B | INT4 Q8 | 51 tok/s 30 tok/s | ~4.4s ~4.8s | |
- | ChatGLM3-6B-32K | 6B | INT4 Q8 | 50 tok/s 36 tok/s | ~3.0s ~3.0s | |
- | Yi-6B | 6B | INT4 Q8 | 50 tok/s 36 tok/s | ~3.0s ~3.0s | |
- | Falcon3-10B | 10B | INT4 Q8 | 47 tok/s 26 tok/s | ~4.8s ~5.7s | |
- | Mochi 1 Preview | 10B | INT4 Q8 | 47 tok/s 26 tok/s | ~4.8s ~5.7s | |
- | Falcon2-11B | 11B | INT4 Q8 | 45 tok/s 24 tok/s | ~5.3s ~6.3s | |
- | Hunyuan Lite | 7B | INT4 Q8 | 45 tok/s 30 tok/s | ~3.4s ~3.7s | |
- | OLMo 3 7B Base | 7B | INT4 Q8 | 45 tok/s 30 tok/s | ~3.4s ~3.7s | |
- | SEA-LION-7B | 7.1B | INT4 Q8 | 44 tok/s 30 tok/s | ~3.5s ~3.8s | |
- | SEA-LION-7B-Instruct | 7.1B | INT4 Q8 | 44 tok/s 30 tok/s | ~3.5s ~3.8s | |
- | FLUX.1 Canny Dev | 11.9B | INT4 | 41 tok/s | ~5.7s | |
- | FLUX.1 Depth Dev | 11.9B | INT4 | 41 tok/s | ~5.7s | |
- | FLUX.1 Dev | 11.9B | INT4 | 41 tok/s | ~5.7s | |
- | FLUX.1 Fill Dev | 11.9B | INT4 | 41 tok/s | ~5.7s | |
- | FLUX.1 Schnell | 11.9B | INT4 | 41 tok/s | ~5.7s | |
- | Sahabat-AI-Llama3-8B-Instruct | 8B | INT4 Q8 | 41 tok/s 26 tok/s | ~3.9s ~4.6s | |
- | Typhoon-2-8B | 8B | INT4 Q8 | 41 tok/s 26 tok/s | ~3.9s ~4.6s | |
- | HunyuanVideo 1.5 720P T2V | 13B | INT4 | 38 tok/s | ~6.3s | |
- | GLM-4-9B | 9B | INT4 | 37 tok/s | ~4.4s | |
- | GLM-4-9B-Chat | 9B | INT4 | 37 tok/s | ~4.4s | |
- | GLM-4-9B-Chat-1M | 9B | INT4 | 37 tok/s | ~4.4s | |
- | Sahabat-AI-Gemma2-9B | 9.2B | INT4 | 37 tok/s | ~4.5s | |
- | Sahabat-AI-Gemma2-9B-Instruct | 9.2B | INT4 | 37 tok/s | ~4.5s | |
- | Yi-9B | 9B | INT4 | 37 tok/s | ~4.4s | |
- | Wan 2.1 I2V 14B 720P | 14B | INT4 | 35 tok/s | ~6.7s | |
- | Wan 2.1 T2V 14B | 14B | INT4 | 35 tok/s | ~6.7s | |
- | Wan 2.2 T2V A14B | 14B | INT4 | 35 tok/s | ~6.7s | |
- | FLUX.2 Dev | 15B | INT4 | 33 tok/s | ~7.2s | |
- | GreenMind-14B-R1 | 14B | INT4 | 23 tok/s | ~7.9s | |
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 AMD RX 6700 XT (12GB).
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 384 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 85 tok/s on an 8B Q4 model and 10 tok/s on a 70B Q4 model.
Frequently Asked Questions
©2025 ApX Machine Learning