NVIDIA · workstation

RTX A4500

RTX A4500 has 20 GB of VRAM at 640 GB/s — about 18.60 GiB usable after driver and compositor overhead. 1029 of 2118 indexed models fit at 128K context with f16 KV.

Spec sheet· bandwidth, theoreticalFrom the file· fit from summed bytesPredicted· speed
Memory
20 GB
GDDR6
Bandwidth
640 GB/s
320-bit bus
Tensor FP16
95 TF
dense
TDP
200 W
Context4K8K16K32K64K128K
KV cachef16q8_0q4_0quantizing the KV cache is a ~2× lever on the dominant term at long context
text 812audio asr 38vision language 121video 16audio tts 20image 1embedding 21

What fits at 128K context

largest quantization that fits, per model · 1029 of 2118 indexed
ModelBest quantParamsWeights●KV●Total in memory◐Headroom◐tok/s◐
Le-Chaton-Slim-23BMoEI1-IQ1_S23.3B4.59 GiB13.00 GiB18.60 GiB0.00 GiB15±37%
Voxtral-Mini-3B-2507Q4_K_L4.7B2.58 GiB15.00 GiB18.59 GiB0.01 GiB21±22%
Llama-3.2-3B-Instruct-abliteratedQ8_03.6B3.58 GiB14.00 GiB18.59 GiB0.01 GiB21±22%
Llama-3.2-3B-Instruct-uncensoredQ8_03.6B3.58 GiB14.00 GiB18.59 GiB0.01 GiB21±22%
dolphincoder-starcoder2-15bKV unresolvedI1-Q3_K_M16.0B7.49 GiB10.00 GiB18.58 GiB0.02 GiB21±22%
gemma-4-12B-coder-fable5-composer2.5-v1-abliteratedQ5_K_M12.0B9.07 GiB8.47 GiB18.58 GiB0.02 GiB21±22%
gemma-4-12B-coder-fable5-composer2.5-v1-sft-v5-abliteratedQ5_K_M12.0B9.07 GiB8.47 GiB18.58 GiB0.02 GiB21±22%
Ling-liteMoEQ4_K_L16.8B10.59 GiB7.00 GiB18.58 GiB0.02 GiB24±37%
Phi-4-mini-instruct-abliteratedQ2_K3.8B1.57 GiB16.00 GiB18.58 GiB0.02 GiB21±22%
Phi-4-mini-reasoningQ2_K3.8B1.57 GiB16.00 GiB18.58 GiB0.02 GiB21±22%
Phi-4-mini-instructQ2_K3.8B1.57 GiB16.00 GiB18.58 GiB0.02 GiB21±22%
Marco-Nano-InstructMoEI1-IQ3_S8.0B3.60 GiB14.00 GiB18.58 GiB0.02 GiB15±37%
gemma-4-26B-A4B-itMoEQ3_K_L26.5B12.29 GiB5.29 GiB18.57 GiB0.03 GiB21±22%
Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16UD-IQ3_S33.0B17.53 GiB0.00 GiB18.57 GiB0.03 GiB21±22%
Qwen3-16B-A3BMoEQ2_K16.0B5.58 GiB12.00 GiB18.57 GiB0.03 GiB16±37%
Pantheon-Reasoning-27BIQ2_XS27.8B9.50 GiB8.00 GiB18.56 GiB0.04 GiB21±22%
Qwen3.5-27BIQ2_XS27.8B9.50 GiB8.00 GiB18.56 GiB0.04 GiB21±22%
Skywork-R1V3-38BQ4_K_S38.4B17.49 GiB0.00 GiB18.56 GiB0.04 GiB21±22%
Qwen3.6-28BMoEI1-Q4_028.2B15.05 GiB2.50 GiB18.55 GiB0.05 GiB52±37%
Qwen3.5-28BMoEI1-Q4_028.7B15.05 GiB2.50 GiB18.55 GiB0.05 GiB52±37%
gemma-4-19B-A4B-it-INSTRUCT-Heretic-UncensoredMoEI1-Q5_K_S19.0B12.27 GiB5.29 GiB18.55 GiB0.05 GiB21±22%
gemma-4-19B-A4B-it-The-DECKARD-Heretic-Uncensored-ThinkingMoEI1-Q5_K_S19.0B12.27 GiB5.29 GiB18.55 GiB0.05 GiB21±22%
gemma-4-19b-a4b-it-REAP-hereticMoEI1-Q5_K_S19.0B12.27 GiB5.29 GiB18.55 GiB0.05 GiB21±22%
Gemma-4-19BMoEI1-Q5_K_S19.0B12.27 GiB5.29 GiB18.55 GiB0.05 GiB21±22%
dolphin-2.9.3-mistral-7B-32kI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-Instruct-v0.3-ParasiteI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-Instruct-v0.3-JbliteratedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-Instruct-v0.3IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mathstral-7B-v0.1IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-v0.3IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
SciPhi-Self-RAG-Mistral-7B-32kKV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
dolphin-2.2.1-mistral-7bKV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
OpenChat-3.5-7B-Qwen-v2.0KV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
openchat-3.5-0106KV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-Instruct-v0.1KV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Mistral-7B-Instruct-v0.2I1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
ContextualKunoichi_KTO-7BI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
xLAM-7b-rI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Ninja-v1-RP-WIPKV unresolvedI1-IQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Kunoichi-DPO-v2-7BKV unresolvedIQ1_S7.2B1.50 GiB16.00 GiB18.54 GiB0.06 GiB21±22%
Aura-4BI1-IQ2_S4.5B1.51 GiB16.00 GiB18.53 GiB0.07 GiB21±22%
magnum-v2-4bI1-IQ2_S4.5B1.51 GiB16.00 GiB18.53 GiB0.07 GiB21±22%
gemma-3-12b-it-vl-Gemini-3-Pro-Preview-Heretic-Uncensored-ThinkingI1-Q6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-vl-Deepseek-v3.1-Heretic-Uncensored-ThinkingI1-Q6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-ultra-uncensored-hereticQ6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-vl-GLM-4.7-Flash-Heretic-Uncensored-ThinkingI1-Q6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
Floppa-12B-Gemma3-UncensoredI1-Q6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-hereticI1-Q6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-abliteratedQ6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-it-abliterated-v2Q6_K11.8B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
gemma-3-12b-itQ6_K12.2B9.00 GiB8.47 GiB18.51 GiB0.09 GiB21±22%
Falcon3-7B-InstructQ3_K_M7.5B3.43 GiB14.00 GiB18.50 GiB0.10 GiB21±22%
OLMoE-1B-7B-0924-InstructMoEI1-IQ1_M6.9B1.53 GiB16.00 GiB18.50 GiB0.10 GiB13±37%
Qwen3.5-27B-Engineer-Deckard-GeminiI1-Q2_K27.7B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Qwen3.5-27B-HERETIC-Polaris-Advanced-Thinking-Alpha-uncensoredI1-Q2_K27.4B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Qwen3.5-27B-Deckard-PKD-Heretic-Uncensored-ThinkingI1-Q2_K27.4B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Huihui-Qwen3.5-27B-abliteratedI1-Q2_K27.8B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Qwen3.5-27B-Unredacted-MAXI1-Q2_K27.4B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Qwen3.5-27B-hereticI1-Q2_K27.4B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
Qwen3.5-27B-DerestrictedI1-Q2_K27.8B9.43 GiB8.00 GiB18.49 GiB0.11 GiB21±22%
●From the file◐Predictedwhat these mean

Speed is modeled, not measured: decode is memory-bandwidth bound, so tokens per second is bytes read per token against achievable bandwidth. Mixture-of-experts models carry a wider band because only the routed experts are read each step, and few have been measured publicly.

Measured on this card

third-party benchmarks, aggregated
Workload◍MedianMiddle 50%Runs
Image generation15.27 it/s11.41–18.32100
Prompt processing2862.73 tok/s2484.22–3293.5014
Text generation95.99 tok/s92.45–96.8012
Benchmarked· n=100

Aggregated from community-submitted runs, so the spread is wide by nature — it covers different models, resolutions, step counts and settings, not one controlled configuration. Read the middle 50% rather than the median alone. These figures are reproduced with attribution from vladmandic-sd-data-benchmark, which publishes no licence — so we display and link rather than redistribute them.

Questions people ask

What AI models can a RTX A4500 run?
1029 of 2118 indexed open-weight models fit a RTX A4500 at 131,072 context with f16 KV cache, the largest being Le-Chaton-Slim-23B at I1-IQ1_S. That covers text, vision-language, image, video and speech models.
How much usable memory does a RTX A4500 actually have?
Its nameplate is 20 GB, but about 18.60 GiB is available to a model once driver and compositor overhead is accounted for.
Is a RTX A4500 fast for local AI?
Its memory bandwidth is 640 GB/s, and that figure — not teraflops — is what governs token generation speed. Capacity decides what you can run; bandwidth decides how fast it runs.