What are the best models to run on NVIDIA RTX 4500 Ada Generation (24GB)?

Based on model rankings and 24 GB usable VRAM: Top Overall: Qwen 3.8 27B (INT4) ~22 tok/s. Top Coding: Qwen3.5-27B (INT4) ~22 tok/s. Optimal Daily Driver: Phi-4 Reasoning Plus (INT4) ~30 tok/s. Maximum Model Size: Falcon-40B (INT4) ~15 tok/s.

How are VRAM, throughput, and latency calculated?

Calculations are determined using apxml.com modeling, combining LLM architecture analysis with GPU hardware physics. VRAM and Context Scaling: Memory requirements are determined from model architecture, including parameter count, layer configuration, and attention structure. The KV cache footprint scales deterministically based on context length, batch size, and cache precision. Throughput (TPS): Token generation is calculated in the memory-bandwidth-bound decode phase, measuring the time required to stream model weights and KV cache relative to hardware memory bandwidth. Latency (TTFT): Time to first token is calculated from the compute operations required during prompt prefill relative to hardware tensor compute capacity.

How accurate are these estimates?

Calculations are calibrated against empirical ground-truth benchmark runs across diverse architectures and hardware. Because the underlying mechanics are benchmark-calibrated, estimates remain accurate across different configurations without needing to run physical tests for every scenario. VRAM Accuracy: Memory estimates are deterministic and closely match actual allocation. Minor differences typically stay within 3% to 5% due to CUDA runtime overhead or memory allocator fragmentation. Speed and Latency Accuracy: Throughput and TTFT typically vary within 5% to 10% depending on the specific inference runtime, kernel implementations, and system power limits.

Best Models to Run on NVIDIA RTX 4500 Ada Generation (24GB) (October 2026)

Rank	Model	Parameters	Precision	TPS (Approx.)	TTFT
#33	Qwen 3.8 27B	27B	INT4	22 tok/s	~1.9s
#48	Agnes 3.0 Flash	33B	INT4	19 tok/s	~2.3s
#72	K2 Horizon MoVA 36B A4B	36B (4B active)	INT4	87 tok/s	~407 ms
#83	Qwen3.5-27B	27B	INT4	22 tok/s	~1.9s
#92	Sarvam-30B	32B (2.4B active)	INT4	119 tok/s	~264 ms
#94	Qwen3.6 35B A3B	35B (3B active)	INT4	102 tok/s	~332 ms
#96	Muse Glimmer 30B	30B	INT4	20 tok/s	~2.1s
#100	Phi-4 Reasoning Plus	14B	INT4 Q8	30 tok/s 20 tok/s	~1.0s ~1.0s
#101	MiMo V2 Flash	15B (309B active)	INT4 Q8	2 tok/s 1 tok/s	~20.7s ~22.8s
#108	Gemma 4 26B A4B	25.2B (3.8B active)	INT4	98 tok/s	~344 ms
#109	Gemma 4 12B	11.95B	INT4 Q8	34 tok/s 23 tok/s	~864 ms ~878 ms
#113	Qwen3.5-9B	9B	INT4 Q8 FP16	59 tok/s 36 tok/s 18 tok/s	~649 ms ~661 ms ~742 ms
#114	Qwen3.5-35B-A3B	35B (3B active)	INT4	102 tok/s	~332 ms
#118	K2 Horizon 7B	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
#128	NVIDIA Nemotron 3 Nano 30B-A3B	3.5B (30B active)	INT4 Q8 FP16	19 tok/s 11 tok/s 6 tok/s	~2.0s ~2.1s ~2.2s
#129	Gemma 4 31B	30.7B	INT4	20 tok/s	~2.1s
#131	Nemotron 3.5 Lightning	30B (3B active)	INT4	40 tok/s	~341 ms
#132	Qwen3.5-4B	4B	INT4 Q8 FP16	108 tok/s 71 tok/s 40 tok/s	~301 ms ~306 ms ~317 ms
#133	Qwen3-30B-A3B	30B (3B active)	INT4	108 tok/s	~301 ms
#136	GPT-OSS 20B	21B (3.6B active)	INT4	47 tok/s	~334 ms
#146	Granite 4.2 8B	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#150	Gemma 4 E4B	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#151	Nemotron 3.5 Lightning 30B A3B	30B (3B active)	INT4	108 tok/s	~301 ms
#153	Qwen3-32B	32B	INT4	19 tok/s	~2.2s
#156	GLM-4	32B	INT4	13 tok/s	~2.6s
#158	GLM-4V	9B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~656 ms ~667 ms ~748 ms
#159	Gemma 3 27B	27B	INT4	22 tok/s	~1.9s
#160	Phi-4	14B	INT4 Q8	40 tok/s 24 tok/s	~999 ms ~1.0s
#162	DeepSeek-R1 32B	32B	INT4	20 tok/s	~2.2s
#164	Qwen2.5-32B	32B	INT4	19 tok/s	~2.2s
#168	Magistral Small	24B	INT4	19 tok/s	~1.7s
#169	OLMo 3 7B Think	7B	INT4 Q8 FP16	52 tok/s 36 tok/s 21 tok/s	~515 ms ~523 ms ~542 ms
#170	OLMo 3.1 32B Instruct	32B	INT4	13 tok/s	~2.6s
#172	DeepSeek-R1 14B	14B	INT4 Q8	42 tok/s 24 tok/s	~998 ms ~1.0s
#173	Ministral 3 14B	14B	INT4 Q8	30 tok/s 20 tok/s	~1.0s ~1.0s
#177	Gemma 3 12B	12B	INT4 Q8 FP16	46 tok/s 27 tok/s 13 tok/s	~859 ms ~874 ms ~1.1s
#180	OLMo 3 32B Base	32B	INT4	13 tok/s	~2.6s
#181	Mistral-Small-2501	24B	INT4	25 tok/s	~1.7s
#182	Ministral 3 8B	8B	INT4 Q8 FP16	47 tok/s 32 tok/s 18 tok/s	~585 ms ~594 ms ~667 ms
#183	Qwen3.5-2B	2B	INT4 Q8 FP16	162 tok/s 116 tok/s 71 tok/s	~163 ms ~166 ms ~171 ms
#184	MiniCPM5-2B	2B	INT4 Q8 FP16	162 tok/s 116 tok/s 71 tok/s	~163 ms ~166 ms ~171 ms
#185	Qwen2.5-14B	14B	INT4 Q8	40 tok/s 24 tok/s	~999 ms ~1.0s
#186	Gemma 2 27B	27B	INT4	22 tok/s	~1.9s
#187	OLMo 3.1 32B Think	32B	INT4	13 tok/s	~2.6s
#188	Gemma 4 E2B	5.1B	INT4 Q8 FP16	92 tok/s 58 tok/s 32 tok/s	~378 ms ~384 ms ~398 ms
#189	Qwen3-4B	4B	INT4 Q8 FP16	108 tok/s 71 tok/s 40 tok/s	~301 ms ~306 ms ~317 ms
#190	DeepSeek-R1 8B	8B	INT4 Q8 FP16	67 tok/s 40 tok/s 21 tok/s	~578 ms ~588 ms ~610 ms
#193	Ministral 3 3B	3B	INT4 Q8 FP16	87 tok/s 67 tok/s 43 tok/s	~236 ms ~239 ms ~247 ms
#195	Qwen3-14B	14B	INT4 Q8	40 tok/s 24 tok/s	~999 ms ~1.0s
#197	Gemma 2 9B	9B	INT4 Q8 FP16	59 tok/s 36 tok/s 18 tok/s	~649 ms ~661 ms ~742 ms
#198	Qwen2.5-7B	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
#199	Qwen3-8B	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#200	Gemma 3 4B	4B	INT4 Q8 FP16	108 tok/s 71 tok/s 40 tok/s	~301 ms ~306 ms ~317 ms
#201	DeepSeek-R1 7B	7B	INT4 Q8 FP16	74 tok/s 45 tok/s 24 tok/s	~509 ms ~518 ms ~537 ms
#202	Llama 3.1 8B	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#204	Qwen3-1.7B	1.7B	INT4 Q8 FP16	177 tok/s 129 tok/s 81 tok/s	~141 ms ~143 ms ~147 ms
#205	Command R	35B	INT4	12 tok/s	~2.9s
#206	Gemma 3 270M	0.27B	INT4 Q8 FP16	285 tok/s 263 tok/s 220 tok/s	~28 ms ~29 ms ~29 ms
#208	Llama 3 8B	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#209	Qwen3-0.6B	0.6B	INT4 Q8 FP16	268 tok/s 224 tok/s 164 tok/s	~54 ms ~55 ms ~56 ms
#210	Phi-4-Mini	3.8B	INT4 Q8 FP16	112 tok/s 74 tok/s 42 tok/s	~287 ms ~292 ms ~302 ms
#212	Ministral-8B-2410	8B	INT4 Q8 FP16	65 tok/s 40 tok/s 21 tok/s	~579 ms ~589 ms ~611 ms
#213	DeepSeek-R1 1.5B	1.5B	INT4 Q8 FP16	188 tok/s 140 tok/s 89 tok/s	~125 ms ~127 ms ~131 ms
#214	Llama 3.2 3B	3B	INT4 Q8 FP16	130 tok/s 88 tok/s 51 tok/s	~232 ms ~236 ms ~244 ms
#215	Qwen3.5-0.8B	0.8B	INT4 Q8 FP16	244 tok/s 197 tok/s 138 tok/s	~70 ms ~71 ms ~74 ms
#216	Phi-3-mini	3.8B	INT4 Q8 FP16	112 tok/s 74 tok/s 42 tok/s	~287 ms ~292 ms ~302 ms
#217	Gemma 3n E2B IT	6B (2B active)	INT4 Q8 FP16	87 tok/s 74 tok/s 53 tok/s	~179 ms ~181 ms ~186 ms
#218	Mistral-7B-Instruct-v0.1	7.3B	INT4 Q8 FP16	70 tok/s 43 tok/s 23 tok/s	~530 ms ~539 ms ~559 ms
#219	Llama 3.2 1B	1B	INT4 Q8 FP16	223 tok/s 176 tok/s 119 tok/s	~88 ms ~89 ms ~92 ms
#220	Gemma 2 2B	2B	INT4 Q8 FP16	162 tok/s 116 tok/s 71 tok/s	~163 ms ~166 ms ~171 ms
#221	Phi-3-medium	14B	INT4 Q8	40 tok/s 24 tok/s	~999 ms ~1.0s
#222	OLMo 3 7B Instruct	7B	INT4 Q8 FP16	52 tok/s 36 tok/s 21 tok/s	~515 ms ~523 ms ~542 ms
#223	Yi-34B	34B	INT4	13 tok/s	~2.8s
#224	Qwen2-7B	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
#225	Phi-3-small	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
#226	Gemma 3 1B	1B	INT4 Q8 FP16	223 tok/s 176 tok/s 119 tok/s	~88 ms ~89 ms ~92 ms
#227	Mistral-7B-Instruct-v0.2	7.3B	INT4 Q8 FP16	70 tok/s 43 tok/s 23 tok/s	~530 ms ~539 ms ~559 ms
#229	Gemma 1 7B	7B	INT4 Q8 FP16	52 tok/s 36 tok/s 21 tok/s	~515 ms ~523 ms ~542 ms
#230	Gemma 1 2B	2B	INT4 Q8 FP16	173 tok/s 121 tok/s 73 tok/s	~163 ms ~165 ms ~171 ms
#231	ChatGLM3-6B	6B	INT4 Q8 FP16	57 tok/s 41 tok/s 25 tok/s	~445 ms ~452 ms ~469 ms
#232	ChatGLM2-6B	6B	INT4 Q8 FP16	57 tok/s 41 tok/s 25 tok/s	~445 ms ~452 ms ~469 ms
#233	ChatGLM-6B	6B	INT4 Q8 FP16	57 tok/s 41 tok/s 25 tok/s	~445 ms ~452 ms ~469 ms
Unranked Models (Sorted by Speed)
-	Qwen2-0.5B	0.5B	INT4 Q8 FP16	282 tok/s 240 tok/s 181 tok/s	~46 ms ~47 ms ~48 ms
-	Qwen2.5-0.5B	0.5B	INT4 Q8 FP16	282 tok/s 240 tok/s 181 tok/s	~46 ms ~47 ms ~48 ms
-	ERNIE-4.5-0.3B	0.3B	INT4 Q8 FP16	277 tok/s 253 tok/s 210 tok/s	~31 ms ~31 ms ~32 ms
-	ERNIE-4.5-0.3B-Base	0.3B	INT4 Q8 FP16	277 tok/s 253 tok/s 210 tok/s	~31 ms ~31 ms ~32 ms
-	Falcon-1B	1B	INT4 Q8 FP16	235 tok/s 182 tok/s 122 tok/s	~87 ms ~89 ms ~91 ms
-	Falcon3-1B	1B	INT4 Q8 FP16	223 tok/s 176 tok/s 119 tok/s	~88 ms ~89 ms ~92 ms
-	Qwen2-1.5B	1.5B	INT4 Q8 FP16	188 tok/s 140 tok/s 89 tok/s	~125 ms ~127 ms ~131 ms
-	Qwen2.5-1.5B	1.5B	INT4 Q8 FP16	188 tok/s 140 tok/s 89 tok/s	~125 ms ~127 ms ~131 ms
-	Spark X2.5 1.7B	1.7B	INT4 Q8 FP16	177 tok/s 129 tok/s 81 tok/s	~141 ms ~143 ms ~147 ms
-	Hy-MT2 1.8B	1.8B	INT4 Q8 FP16	172 tok/s 125 tok/s 77 tok/s	~148 ms ~150 ms ~155 ms
-	Falcon-3B	3B	INT4 Q8 FP16	138 tok/s 91 tok/s 52 tok/s	~232 ms ~235 ms ~244 ms
-	CroissantLLM Base	1.3B	INT4 Q8 FP16	135 tok/s 114 tok/s 81 tok/s	~112 ms ~114 ms ~117 ms
-	Phi-1	1.3B	INT4 Q8 FP16	135 tok/s 114 tok/s 81 tok/s	~112 ms ~114 ms ~117 ms
-	Phi-1.5	1.3B	INT4 Q8 FP16	135 tok/s 114 tok/s 81 tok/s	~112 ms ~114 ms ~117 ms
-	DeepSeek-R1 3B	3B	INT4 Q8 FP16	132 tok/s 89 tok/s 51 tok/s	~232 ms ~236 ms ~244 ms
-	Falcon3-3B	3B	INT4 Q8 FP16	130 tok/s 88 tok/s 51 tok/s	~232 ms ~236 ms ~244 ms
-	Qwen2.5-3B	3B	INT4 Q8 FP16	130 tok/s 88 tok/s 51 tok/s	~232 ms ~236 ms ~244 ms
-	ERNIE-4.5-21B-A3B	21B (3B active)	INT4 Q8	114 tok/s 73 tok/s	~282 ms ~334 ms
-	ERNIE-4.5-21B-A3B-Base	21B (3B active)	INT4 Q8	114 tok/s 73 tok/s	~282 ms ~334 ms
-	ERNIE-4.5-VL-28B-A3B	28B (3B active)	INT4	110 tok/s	~296 ms
-	ERNIE-4.5-VL-28B-A3B-Base	28B (3B active)	INT4	110 tok/s	~296 ms
-	Hy-MT2 30B A3B	30B (3B active)	INT4	108 tok/s	~301 ms
-	Spark X2.5 4B	4B	INT4 Q8 FP16	108 tok/s 71 tok/s 40 tok/s	~301 ms ~306 ms ~317 ms
-	Phi-2	2.7B	INT4 Q8 FP16	91 tok/s 72 tok/s 47 tok/s	~215 ms ~218 ms ~225 ms
-	MaLLaM-3B	3B	INT4 Q8 FP16	87 tok/s 67 tok/s 43 tok/s	~236 ms ~239 ms ~247 ms
-	SmolLM3 3B	3B	INT4 Q8 FP16	87 tok/s 67 tok/s 43 tok/s	~236 ms ~239 ms ~247 ms
-	Falcon-7B	7B	INT4 Q8 FP16	75 tok/s 46 tok/s 24 tok/s	~509 ms ~518 ms ~537 ms
-	Falcon3-7B	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
-	Hy-MT2 7B	7B	INT4 Q8 FP16	72 tok/s 44 tok/s 24 tok/s	~510 ms ~518 ms ~537 ms
-	Mistral-7B-v0.1	7.3B	INT4 Q8 FP16	70 tok/s 43 tok/s 23 tok/s	~530 ms ~539 ms ~559 ms
-	LensVLM 9B	9B	INT4 Q8 FP16	59 tok/s 36 tok/s 18 tok/s	~649 ms ~661 ms ~742 ms
-	MiMo V2.6 Distill Qwen 9B	9B	INT4 Q8 FP16	59 tok/s 36 tok/s 18 tok/s	~649 ms ~661 ms ~742 ms
-	ChatGLM3-6B-32K	6B	INT4 Q8 FP16	57 tok/s 41 tok/s 25 tok/s	~445 ms ~452 ms ~469 ms
-	Yi-6B	6B	INT4 Q8 FP16	57 tok/s 41 tok/s 25 tok/s	~445 ms ~452 ms ~469 ms
-	Falcon3-10B	10B	INT4 Q8 FP16	54 tok/s 32 tok/s 16 tok/s	~719 ms ~731 ms ~822 ms
-	Kimi-VL-A3B-Instruct	16B (3B active)	INT4 Q8	54 tok/s 45 tok/s	~281 ms ~307 ms
-	Falcon2-11B	11B	INT4 Q8 FP16	52 tok/s 30 tok/s 14 tok/s	~788 ms ~801 ms ~972 ms
-	Hunyuan Lite	7B	INT4 Q8 FP16	52 tok/s 36 tok/s 21 tok/s	~515 ms ~523 ms ~542 ms
-	OLMo 3 7B Base	7B	INT4 Q8 FP16	52 tok/s 36 tok/s 21 tok/s	~515 ms ~523 ms ~542 ms
-	SEA-LION-7B	7.1B	INT4 Q8 FP16	51 tok/s 36 tok/s 21 tok/s	~522 ms ~530 ms ~550 ms
-	SEA-LION-7B-Instruct	7.1B	INT4 Q8 FP16	51 tok/s 36 tok/s 21 tok/s	~522 ms ~530 ms ~550 ms
-	Sahabat-AI-Llama3-8B-Instruct	8B	INT4 Q8 FP16	47 tok/s 32 tok/s 18 tok/s	~585 ms ~594 ms ~667 ms
-	Typhoon-2-8B	8B	INT4 Q8 FP16	47 tok/s 32 tok/s 18 tok/s	~585 ms ~594 ms ~667 ms
-	GLM-4-9B	9B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~656 ms ~667 ms ~748 ms
-	GLM-4-9B-Chat	9B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~656 ms ~667 ms ~748 ms
-	GLM-4-9B-Chat-1M	9B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~656 ms ~667 ms ~748 ms
-	Sahabat-AI-Gemma2-9B	9.2B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~670 ms ~680 ms ~764 ms
-	Sahabat-AI-Gemma2-9B-Instruct	9.2B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~670 ms ~680 ms ~764 ms
-	Yi-9B	9B	INT4 Q8 FP16	42 tok/s 29 tok/s 16 tok/s	~656 ms ~667 ms ~748 ms
-	GreenMind-14B-R1	14B	INT4 Q8	30 tok/s 20 tok/s	~1.0s ~1.0s
-	Falcon-40B	40B	INT4	15 tok/s	~3.0s

NVIDIA RTX 4500 Ada Generation (24GB)

VRAM Scaling by Context Length

Best Models to Run on NVIDIA RTX 4500 Ada Generation (24GB)

Inference Capacity

Fine-Tuning Capacity

Workload Recommendations

Frequently Asked Questions