NVIDIA RTX 4090 D (24GB)
The NVIDIA RTX 4090 D (24GB) features 24 GB of dedicated VRAM and 1008 GB/s memory bandwidth, supporting CUDA and TensorRT-LLM runtimes for full-precision and quantized inference.
Memory
24 GB
Usable Memory Ceiling
24 GB
Bandwidth
1008 GB/s
Max Inference Class
32B Models (Q4)
TDP
425W
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 NVIDIA RTX 4090 D (24GB)
Inference throughput, memory compatibility, and generation speeds for open-weights LLMs on NVIDIA RTX 4090 D (24GB).
Sort by:
Context:
KV Cache:
GPUs:
| Rank | Model | Parameters | Precision | TPS (Approx.) | TTFT | Can Run |
|---|---|---|---|---|---|---|
#36 | Qwen 3.8 27B | 27B | INT4 | 48 tok/s | ~2.0s | |
#80 | K2 Horizon MoVA 36B A4B | 36B (4B active) | INT4 | 153 tok/s | ~433 ms | |
#86 | Qwen3.5-27B | 27B | INT4 | 48 tok/s | ~2.0s | |
#93 | Sarvam-30B | 32B (2.4B active) | INT4 | 194 tok/s | ~280 ms | |
#99 | Muse Glimmer 30B | 30B | INT4 | 44 tok/s | ~2.2s | |
#102 | Phi-4 Reasoning Plus | 14B | INT4 Q8 | 62 tok/s 43 tok/s | ~1.1s ~1.1s | |
#103 | Qwen3.6 35B A3B | 35B (3B active) | INT4 | 171 tok/s | ~353 ms | |
#105 | MiMo V2 Flash | 15B (309B active) | INT4 Q8 | 5 tok/s 3 tok/s | ~22.0s ~24.0s | |
#110 | Gemma 4 26B A4B | 25.2B (3.8B active) | INT4 | 169 tok/s | ~365 ms | |
#120 | Qwen3.5-35B-A3B | 35B (3B active) | INT4 | 171 tok/s | ~353 ms | |
#121 | Qwen3.5-9B | 9B | INT4 Q8 FP16 | 113 tok/s 74 tok/s 39 tok/s | ~690 ms ~694 ms ~764 ms | |
#127 | K2 Horizon 7B | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
#130 | Gemma 4 12B | 11.95B | INT4 Q8 | 70 tok/s 49 tok/s | ~913 ms ~919 ms | |
#136 | Qwen3.5-4B | 4B | INT4 Q8 FP16 | 182 tok/s 131 tok/s 81 tok/s | ~320 ms ~322 ms ~327 ms | |
#139 | GPT-OSS 20B | 21B (3.6B active) | INT4 | 93 tok/s | ~347 ms | |
#140 | GLM-4V | 9B | INT4 Q8 FP16 | 86 tok/s 62 tok/s 36 tok/s | ~692 ms ~697 ms ~767 ms | |
#145 | NVIDIA Nemotron 3 Nano 30B-A3B | 3.5B (30B active) | INT4 Q8 FP16 | 42 tok/s 25 tok/s 13 tok/s | ~2.2s ~2.2s ~2.2s | |
#150 | Qwen3-30B-A3B | 30B (3B active) | INT4 | 182 tok/s | ~319 ms | |
#152 | Granite 4.2 8B | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#157 | Gemma 4 E4B | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#161 | GLM-4 | 32B | INT4 | 30 tok/s | ~2.8s | |
#163 | Gemma 3 27B | 27B | INT4 | 48 tok/s | ~2.0s | |
#164 | Phi-4 | 14B | INT4 Q8 | 82 tok/s 51 tok/s | ~1.1s ~1.1s | |
#166 | Qwen3-32B | 32B | INT4 | 42 tok/s | ~2.4s | |
#168 | DeepSeek-R1 32B | 32B | INT4 | 43 tok/s | ~2.4s | |
#169 | Gemma 4 31B | 30.7B | INT4 | 43 tok/s | ~2.3s | |
#171 | Magistral Small | 24B | INT4 | 41 tok/s | ~1.8s | |
#172 | Qwen2.5-32B | 32B | INT4 | 42 tok/s | ~2.4s | |
#174 | OLMo 3.1 32B Instruct | 32B | INT4 | 30 tok/s | ~2.8s | |
#176 | DeepSeek-R1 14B | 14B | INT4 Q8 | 85 tok/s 52 tok/s | ~1.1s ~1.1s | |
#177 | Ministral 3 14B | 14B | INT4 Q8 | 62 tok/s 43 tok/s | ~1.1s ~1.1s | |
#180 | Gemma 3 12B | 12B | INT4 Q8 FP16 | 92 tok/s 58 tok/s 29 tok/s | ~913 ms ~919 ms ~1.1s | |
#184 | Mistral-Small-2501 | 24B | INT4 | 53 tok/s | ~1.8s | |
#185 | Ministral 3 8B | 8B | INT4 Q8 FP16 | 93 tok/s 68 tok/s 39 tok/s | ~617 ms ~621 ms ~684 ms | |
#186 | Gemma 2 27B | 27B | INT4 | 48 tok/s | ~2.0s | |
#187 | OLMo 3 7B Think | 7B | INT4 Q8 FP16 | 101 tok/s 75 tok/s 46 tok/s | ~543 ms ~547 ms ~555 ms | |
#188 | Qwen3.5-2B | 2B | INT4 Q8 FP16 | 239 tok/s 191 tok/s 132 tok/s | ~173 ms ~174 ms ~176 ms | |
#189 | MiniCPM5-2B | 2B | INT4 Q8 FP16 | 239 tok/s 191 tok/s 132 tok/s | ~173 ms ~174 ms ~176 ms | |
#190 | Qwen2.5-14B | 14B | INT4 Q8 | 82 tok/s 51 tok/s | ~1.1s ~1.1s | |
#191 | OLMo 3.1 32B Think | 32B | INT4 | 30 tok/s | ~2.8s | |
#193 | Gemma 4 E2B | 5.1B | INT4 Q8 FP16 | 161 tok/s 112 tok/s 67 tok/s | ~401 ms ~404 ms ~410 ms | |
#195 | Ministral 3 3B | 3B | INT4 Q8 FP16 | 154 tok/s 126 tok/s 88 tok/s | ~248 ms ~249 ms ~253 ms | |
#196 | DeepSeek-R1 8B | 8B | INT4 Q8 FP16 | 125 tok/s 82 tok/s 46 tok/s | ~614 ms ~619 ms ~628 ms | |
#198 | Gemma 2 9B | 9B | INT4 Q8 FP16 | 113 tok/s 74 tok/s 39 tok/s | ~690 ms ~694 ms ~764 ms | |
#199 | Nemotron 3.5 Lightning 30B A3B | 30B (3B active) | INT4 | 182 tok/s | ~319 ms | |
#201 | Qwen3-4B | 4B | INT4 Q8 FP16 | 182 tok/s 131 tok/s 81 tok/s | ~320 ms ~322 ms ~327 ms | |
#202 | Qwen3-14B | 14B | INT4 Q8 | 82 tok/s 51 tok/s | ~1.1s ~1.1s | |
#203 | Gemma 3 4B | 4B | INT4 Q8 FP16 | 182 tok/s 131 tok/s 81 tok/s | ~320 ms ~322 ms ~327 ms | |
#204 | DeepSeek-R1 7B | 7B | INT4 Q8 FP16 | 136 tok/s 91 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
#205 | OLMo 3 32B Base | 32B | INT4 | 30 tok/s | ~2.8s | |
#206 | Qwen3.5-0.8B | 0.8B | INT4 Q8 FP16 | 303 tok/s 269 tok/s 215 tok/s | ~74 ms ~75 ms ~76 ms | |
#207 | Llama 3.1 8B | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#207 | Qwen3-8B | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#210 | Nemotron 3.5 Lightning | 30B (3B active) | INT4 | 81 tok/s | ~353 ms | |
#211 | Command R | 35B | INT4 | 28 tok/s | ~3.0s | |
#212 | Llama 3 8B | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#213 | Qwen3-1.7B | 1.7B | INT4 Q8 FP16 | 252 tok/s 206 tok/s 146 tok/s | ~149 ms ~150 ms ~152 ms | |
#214 | Gemma 3 270M | 0.27B | INT4 Q8 FP16 | 328 tok/s 315 tok/s 287 tok/s | ~29 ms ~30 ms ~30 ms | |
#216 | Qwen3-0.6B | 0.6B | INT4 Q8 FP16 | 318 tok/s 289 tok/s 241 tok/s | ~57 ms ~57 ms ~58 ms | |
#217 | Phi-4-Mini | 3.8B | INT4 Q8 FP16 | 186 tok/s 136 tok/s 85 tok/s | ~305 ms ~307 ms ~311 ms | |
#218 | Qwen2.5-7B | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
#220 | Ministral-8B-2410 | 8B | INT4 Q8 FP16 | 123 tok/s 81 tok/s 46 tok/s | ~615 ms ~619 ms ~628 ms | |
#221 | DeepSeek-R1 1.5B | 1.5B | INT4 Q8 FP16 | 262 tok/s 217 tok/s 157 tok/s | ~132 ms ~133 ms ~135 ms | |
#222 | Llama 3.2 3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
#223 | Phi-3-mini | 3.8B | INT4 Q8 FP16 | 186 tok/s 136 tok/s 85 tok/s | ~305 ms ~307 ms ~311 ms | |
#224 | Gemma 3n E2B IT | 6B (2B active) | INT4 Q8 FP16 | 154 tok/s 136 tok/s 103 tok/s | ~186 ms ~187 ms ~189 ms | |
#225 | Gemma 2 2B | 2B | INT4 Q8 FP16 | 239 tok/s 191 tok/s 132 tok/s | ~173 ms ~174 ms ~176 ms | |
#226 | Mistral-7B-Instruct-v0.1 | 7.3B | INT4 Q8 FP16 | 130 tok/s 87 tok/s 50 tok/s | ~563 ms ~567 ms ~575 ms | |
#227 | Llama 3.2 1B | 1B | INT4 Q8 FP16 | 289 tok/s 251 tok/s 194 tok/s | ~93 ms ~93 ms ~94 ms | |
#228 | Phi-3-medium | 14B | INT4 Q8 | 82 tok/s 51 tok/s | ~1.1s ~1.1s | |
#229 | Yi-34B | 34B | INT4 | 28 tok/s | ~3.0s | |
#230 | OLMo 3 7B Instruct | 7B | INT4 Q8 FP16 | 101 tok/s 75 tok/s 46 tok/s | ~543 ms ~547 ms ~555 ms | |
#231 | Qwen2-7B | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
#232 | Phi-3-small | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
#233 | Mistral-7B-Instruct-v0.2 | 7.3B | INT4 Q8 FP16 | 130 tok/s 87 tok/s 50 tok/s | ~563 ms ~567 ms ~575 ms | |
#234 | Gemma 3 1B | 1B | INT4 Q8 FP16 | 289 tok/s 251 tok/s 194 tok/s | ~93 ms ~93 ms ~94 ms | |
#236 | Gemma 1 7B | 7B | INT4 Q8 FP16 | 101 tok/s 75 tok/s 46 tok/s | ~543 ms ~547 ms ~555 ms | |
#237 | Gemma 1 2B | 2B | INT4 Q8 FP16 | 249 tok/s 197 tok/s 135 tok/s | ~173 ms ~174 ms ~176 ms | |
#238 | ChatGLM3-6B | 6B | INT4 Q8 FP16 | 111 tok/s 83 tok/s 53 tok/s | ~469 ms ~472 ms ~479 ms | |
#239 | ChatGLM2-6B | 6B | INT4 Q8 FP16 | 111 tok/s 83 tok/s 53 tok/s | ~469 ms ~472 ms ~479 ms | |
#240 | ChatGLM-6B | 6B | INT4 Q8 FP16 | 111 tok/s 83 tok/s 53 tok/s | ~469 ms ~472 ms ~479 ms | |
Unranked Models (Sorted by Speed) | ||||||
- | DINOv3 ViT-L/16 | 0.3B | INT4 Q8 FP16 | 343 tok/s 325 tok/s 292 tok/s | ~32 ms ~32 ms ~32 ms | |
- | Qwen2-0.5B | 0.5B | INT4 Q8 FP16 | 326 tok/s 301 tok/s 256 tok/s | ~48 ms ~48 ms ~49 ms | |
- | Qwen2.5-0.5B | 0.5B | INT4 Q8 FP16 | 326 tok/s 301 tok/s 256 tok/s | ~48 ms ~48 ms ~49 ms | |
- | ERNIE-4.5-0.3B | 0.3B | INT4 Q8 FP16 | 323 tok/s 309 tok/s 279 tok/s | ~32 ms ~32 ms ~32 ms | |
- | ERNIE-4.5-0.3B-Base | 0.3B | INT4 Q8 FP16 | 323 tok/s 309 tok/s 279 tok/s | ~32 ms ~32 ms ~32 ms | |
- | Liquid D1-Omni 600M | 0.6B | INT4 Q8 FP16 | 318 tok/s 289 tok/s 241 tok/s | ~57 ms ~57 ms ~58 ms | |
- | Falcon-1B | 1B | INT4 Q8 FP16 | 297 tok/s 257 tok/s 198 tok/s | ~93 ms ~93 ms ~94 ms | |
- | Falcon3-1B | 1B | INT4 Q8 FP16 | 289 tok/s 251 tok/s 194 tok/s | ~93 ms ~93 ms ~94 ms | |
- | MiniCPM-V 4.7 1B | 1B | INT4 Q8 FP16 | 289 tok/s 251 tok/s 194 tok/s | ~93 ms ~93 ms ~94 ms | |
- | Qwen2-1.5B | 1.5B | INT4 Q8 FP16 | 262 tok/s 217 tok/s 157 tok/s | ~132 ms ~133 ms ~135 ms | |
- | Qwen2.5-1.5B | 1.5B | INT4 Q8 FP16 | 262 tok/s 217 tok/s 157 tok/s | ~132 ms ~133 ms ~135 ms | |
- | Spark X2.5 1.7B | 1.7B | INT4 Q8 FP16 | 252 tok/s 206 tok/s 146 tok/s | ~149 ms ~150 ms ~152 ms | |
- | Hy-MT2 1.8B | 1.8B | INT4 Q8 FP16 | 248 tok/s 201 tok/s 141 tok/s | ~156 ms ~157 ms ~159 ms | |
- | Falcon-3B | 3B | INT4 Q8 FP16 | 215 tok/s 160 tok/s 103 tok/s | ~246 ms ~247 ms ~251 ms | |
- | CroissantLLM Base | 1.3B | INT4 Q8 FP16 | 212 tok/s 188 tok/s 146 tok/s | ~117 ms ~118 ms ~119 ms | |
- | Phi-1 | 1.3B | INT4 Q8 FP16 | 212 tok/s 188 tok/s 146 tok/s | ~117 ms ~118 ms ~119 ms | |
- | Phi-1.5 | 1.3B | INT4 Q8 FP16 | 212 tok/s 188 tok/s 146 tok/s | ~117 ms ~118 ms ~119 ms | |
- | DeepSeek-R1 3B | 3B | INT4 Q8 FP16 | 209 tok/s 157 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | Mellum 2.1 12B A2.5B Thinking | 12B (2.5B active) | INT4 Q8 FP16 | 209 tok/s 164 tok/s 100 tok/s | ~240 ms ~242 ms ~286 ms | |
- | Falcon3-3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | LFM2-VL 3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | Liquid D1 3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | Liquid D1-VL 3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | Liquid LFM2-VL 3B D1 | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | Qwen2.5-3B | 3B | INT4 Q8 FP16 | 207 tok/s 156 tok/s 101 tok/s | ~246 ms ~248 ms ~251 ms | |
- | LFM2.5 24B A2B | 24B (2.5B active) | INT4 | 197 tok/s | ~272 ms | |
- | ERNIE-4.5-21B-A3B | 21B (3B active) | INT4 Q8 | 188 tok/s 131 tok/s | ~299 ms ~352 ms | |
- | ERNIE-4.5-21B-A3B-Base | 21B (3B active) | INT4 Q8 | 188 tok/s 131 tok/s | ~299 ms ~352 ms | |
- | ERNIE-4.5-VL-28B-A3B | 28B (3B active) | INT4 | 183 tok/s | ~314 ms | |
- | ERNIE-4.5-VL-28B-A3B-Base | 28B (3B active) | INT4 | 183 tok/s | ~314 ms | |
- | FrogNano 4B | 4B | INT4 Q8 FP16 | 182 tok/s 131 tok/s 81 tok/s | ~320 ms ~322 ms ~327 ms | |
- | Hy-Memory 30B A3B | 30B (3B active) | INT4 | 182 tok/s | ~319 ms | |
- | Hy-MT2 30B A3B | 30B (3B active) | INT4 | 182 tok/s | ~319 ms | |
- | Spark X2.5 4B | 4B | INT4 Q8 FP16 | 182 tok/s 131 tok/s 81 tok/s | ~320 ms ~322 ms ~327 ms | |
- | EmbeddingGemma 2 | 2B | INT4 Q8 FP16 | 176 tok/s 151 tok/s 112 tok/s | ~174 ms ~175 ms ~178 ms | |
- | MiniCPM-V 4.7 35B A3B | 35B (3B active) | INT4 | 171 tok/s | ~353 ms | |
- | Phi-2 | 2.7B | INT4 Q8 FP16 | 160 tok/s 133 tok/s 94 tok/s | ~226 ms ~227 ms ~230 ms | |
- | MaLLaM-3B | 3B | INT4 Q8 FP16 | 154 tok/s 126 tok/s 88 tok/s | ~248 ms ~249 ms ~253 ms | |
- | SmolLM3 3B | 3B | INT4 Q8 FP16 | 154 tok/s 126 tok/s 88 tok/s | ~248 ms ~249 ms ~253 ms | |
- | Falcon-7B | 7B | INT4 Q8 FP16 | 138 tok/s 91 tok/s 52 tok/s | ~541 ms ~544 ms ~553 ms | |
- | Falcon3-7B | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
- | Hy-MT2 7B | 7B | INT4 Q8 FP16 | 134 tok/s 89 tok/s 52 tok/s | ~541 ms ~545 ms ~553 ms | |
- | Mistral-7B-v0.1 | 7.3B | INT4 Q8 FP16 | 130 tok/s 87 tok/s 50 tok/s | ~563 ms ~567 ms ~575 ms | |
- | Youtu Parsing Omni | 8B | INT4 Q8 FP16 | 125 tok/s 82 tok/s 46 tok/s | ~614 ms ~619 ms ~628 ms | |
- | LensVLM 9B | 9B | INT4 Q8 FP16 | 113 tok/s 74 tok/s 39 tok/s | ~690 ms ~694 ms ~764 ms | |
- | MiMo V2.6 Distill Qwen 9B | 9B | INT4 Q8 FP16 | 113 tok/s 74 tok/s 39 tok/s | ~690 ms ~694 ms ~764 ms | |
- | ChatGLM3-6B-32K | 6B | INT4 Q8 FP16 | 111 tok/s 83 tok/s 53 tok/s | ~469 ms ~472 ms ~479 ms | |
- | Yi-6B | 6B | INT4 Q8 FP16 | 111 tok/s 83 tok/s 53 tok/s | ~469 ms ~472 ms ~479 ms | |
- | Falcon3-10B | 10B | INT4 Q8 FP16 | 106 tok/s 68 tok/s 36 tok/s | ~763 ms ~769 ms ~846 ms | |
- | Kimi-VL-A3B-Instruct | 16B (3B active) | INT4 Q8 | 106 tok/s 89 tok/s | ~292 ms ~317 ms | |
- | Falcon2-11B | 11B | INT4 Q8 FP16 | 102 tok/s 64 tok/s 32 tok/s | ~837 ms ~843 ms ~1.0s | |
- | Hunyuan Lite | 7B | INT4 Q8 FP16 | 101 tok/s 75 tok/s 46 tok/s | ~543 ms ~547 ms ~555 ms | |
- | OLMo 3 7B Base | 7B | INT4 Q8 FP16 | 101 tok/s 75 tok/s 46 tok/s | ~543 ms ~547 ms ~555 ms | |
- | SEA-LION-7B | 7.1B | INT4 Q8 FP16 | 101 tok/s 74 tok/s 46 tok/s | ~551 ms ~554 ms ~562 ms | |
- | SEA-LION-7B-Instruct | 7.1B | INT4 Q8 FP16 | 101 tok/s 74 tok/s 46 tok/s | ~551 ms ~554 ms ~562 ms | |
- | Sahabat-AI-Llama3-8B-Instruct | 8B | INT4 Q8 FP16 | 93 tok/s 68 tok/s 39 tok/s | ~617 ms ~621 ms ~684 ms | |
- | Typhoon-2-8B | 8B | INT4 Q8 FP16 | 93 tok/s 68 tok/s 39 tok/s | ~617 ms ~621 ms ~684 ms | |
- | GLM-4-9B | 9B | INT4 Q8 FP16 | 86 tok/s 62 tok/s 36 tok/s | ~692 ms ~697 ms ~767 ms | |
- | GLM-4-9B-Chat | 9B | INT4 Q8 FP16 | 86 tok/s 62 tok/s 36 tok/s | ~692 ms ~697 ms ~767 ms | |
- | GLM-4-9B-Chat-1M | 9B | INT4 Q8 FP16 | 86 tok/s 62 tok/s 36 tok/s | ~692 ms ~697 ms ~767 ms | |
- | Yi-9B | 9B | INT4 Q8 FP16 | 86 tok/s 62 tok/s 36 tok/s | ~692 ms ~697 ms ~767 ms | |
- | Sahabat-AI-Gemma2-9B | 9.2B | INT4 Q8 FP16 | 85 tok/s 61 tok/s 35 tok/s | ~707 ms ~712 ms ~783 ms | |
- | Sahabat-AI-Gemma2-9B-Instruct | 9.2B | INT4 Q8 FP16 | 85 tok/s 61 tok/s 35 tok/s | ~707 ms ~712 ms ~783 ms | |
- | GreenMind-14B-R1 | 14B | INT4 Q8 | 62 tok/s 43 tok/s | ~1.1s ~1.1s | |
- | AREX-2 | 27B | INT4 | 48 tok/s | ~2.0s | |
- | PPLX Decider v1 27B | 27B | INT4 | 48 tok/s | ~2.0s | |
- | PPLX Decider v1.1 27B | 27B | INT4 | 48 tok/s | ~2.0s | |
- | Agnes 3.0 Flash | 33B | INT4 | 41 tok/s | ~2.4s | |
- | Falcon-40B | 40B | INT4 | 34 tok/s | ~3.2s | |
Inference Capacity
Maximum model parameter class runnable by weight precision.
Context:
4-bit
32B Models
8-bit
14B Models
16-bit
8B 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 NVIDIA RTX 4090 D (24GB).
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 1008 GB/s aggregate bandwidth, batch size 1 inference operates in a memory-bandwidth bound regime. Generates approximately 224 tok/s on an 8B Q4 model and 27 tok/s on a 70B Q4 model.
Frequently Asked Questions
©2025 ApX Machine Learning