RTX 4080 SUPER · 16 GB VRAM
NVIDIA · RTX 4080 SUPER
125 text models run well · 736 GB/s bandwidth
Models on this hardware
Start here
File Q4_K_M · 0.8 GB · bartowski/Llama-3.2-1B-Instruct-GGUF · What is Q4?
Memory Needs ~1 GB · ~3 GB on RTX 4080 SUPER · 16 GB VRAM, including ~2 GB the system keeps
ollama run hf.co/bartowski/Llama-3.2-1B-Instruct-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
File Q4_K_M · 2.2 GB · ibm-granite/granite-4.2-3b-GGUF · What is Q4?
Memory Needs ~3 GB · ~5 GB on RTX 4080 SUPER · 16 GB VRAM, including ~2 GB the system keeps
ollama run hf.co/ibm-granite/granite-4.2-3b-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
File Q4_K_M · 5.3 GB · ibm-granite/granite-4.2-8b-GGUF · What is Q4?
Memory Needs ~6 GB · ~8 GB on RTX 4080 SUPER · 16 GB VRAM, including ~2 GB the system keeps
ollama run hf.co/ibm-granite/granite-4.2-8b-GGUF:Q4_K_M- Install Ollama
- Run the command above
- Ask:
Explain what a context window is, in two sentences.
Text
Text · LMArena
rating · higher is better
- Color = publisher
Source: LMArena 2026-10-22 · collected 2026-10-25 · dataset · CC BY 4.0 · Models without a rating are left out, not scored as zero.
| Granite 4.2 3BQ4 · published params · source | 1,290 ± 12 | ~5 GB | Q4 · published params |
|---|---|---|---|
| Granite 4.2 8BQ4 · published params · source | 1,290 ± 11 | ~8 GB | Q4 · published params |
| Phi-4Q4 · published params · source | 1,256 ± 5 | ~13 GB | Q4 · published params |
| Llama 3.1 8B InstructQ4 · published params · source | 1,211 ± 4 | ~8 GB | Q4 · published params |
| Llama 3.2 3B InstructQ4 · published params · source | 1,167 ± 8 | ~5 GB | Q4 · published params |
| Llama 3.2 1B InstructQ4 · published params · source | 1,111 ± 8 | ~3 GB | Q4 · published params |
| Qwen3 14BQ4 · published params · source | Not on the LMArena text leaderboard of 2026-10-22 | ~14 GB | Q4 · published params |
| Qwen3 8BQ4 · published params · source | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | Q4 · published params |
No model within 16 GB has a LiveBench score, so the table has no LiveBench column. LMArena ratings in this table: LMArena 2026-10-22 · collected 2026-10-25 · dataset · CC BY 4.0
Every reviewed text model here 37
Every one that runs here, well or slowly, scored or not (21 run well, 16 slowly); 19 of 37 have a LiveBench or LMArena result. Run well first, then tight fits, then slow; within each, by Hub downloads over 30 days, read 2026-10-25, not by quality.
| Qwen3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~14 GB | 9.0 | Apache 2.0 | 10.4M |
|---|---|---|---|---|---|---|
| Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 5.0 | Apache 2.0 | 8.1M |
| Qwen3.5 9B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 5.7 | Apache 2.0 | 6.1M |
| Llama 3.1 8B Instruct | Not in LiveBench 2026-06-25 | 1,211 | ~8 GB | 4.9 | Llama 3.1 | 4.8M |
| Qwen3 1.7B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~4 GB | 1.1 | Apache 2.0 | 4.1M |
| Gemma 4 E4B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~8 GB | 5.0 | Apache 2.0 | 4M |
| Llama 3.2 1B Instruct | Not in LiveBench 2026-06-25 | 1,111 | ~3 GB | 0.8 | Llama 3.2 | 3.5M |
| Gemma 4 E2B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~6 GB | 3.1 | Apache 2.0 | 3M |
| Qwen3 4B Instruct 2507 | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~5 GB | 2.5 | Apache 2.0 | 2.6M |
| Gemma 4 12B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~11 GB | 7.1 | Apache 2.0 | 1.7M |
| Llama 3.2 3B Instruct | Not in LiveBench 2026-06-25 | 1,167 | ~5 GB | 2.0 | Llama 3.2 | 1.1M |
| SmolLM3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~5 GB | 1.9 | Apache 2.0 | 718.7K |
| Phi-4 | Not in LiveBench 2026-06-25 | 1,256 | ~13 GB | 8.0 | MIT | 368.3K |
| DeepSeek R1 0528 Qwen3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 5.0 | MIT | 355.8K |
| Phi-4 mini | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~5 GB | 2.5 | MIT | 342.8K |
| Olmo 3 7B Instruct | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~8 GB | 4.5 | Apache 2.0 | 262.5K |
| Ministral 3 14B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~13 GB | 8.2 | Apache 2.0 | 126.5K |
| Ministral 3 3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~5 GB | 2.1 | Apache 2.0 | 87.4K |
| Ministral 3 8B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~9 GB | 5.2 | Apache 2.0 | 79.4K |
| Granite 4.2 8B | Not in LiveBench 2026-06-25 | 1,290 | ~8 GB | 5.3 | Apache 2.0 | 78.3K |
| Granite 4.2 3B | Not in LiveBench 2026-06-25 | 1,290 | ~5 GB | 2.2 | Apache 2.0 | 39.8K |
| Gemma 4 31B | Not in LiveBench 2026-06-25 | 1,452 | ~26 GBRuns slowly | 18.3 | Apache 2.0 | 8.4M |
| Gemma 4 26B A4B | Not in LiveBench 2026-06-25 | 1,437 | ~21 GBRuns slowly | 16.9 | Apache 2.0 | 8.2M |
| Qwen3.8 27B | 75.3 | 1,437 | ~23 GBRuns slowly | 16.5 | Apache 2.0 | 7.8M |
| gpt-oss-20b | Not in LiveBench 2026-06-25 | 1,317 | ~18 GBRuns slowly | 11.6 | Apache 2.0 | 5.8M |
| Qwen3.6 27B | 64.0 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GBRuns slowly | 16.8 | Apache 2.0 | 2.4M |
| Qwen3.5 35B A3B | Not in LiveBench 2026-06-25 | 1,394 | ~29 GBRuns slowly | 22.0 | Apache 2.0 | 1.5M |
| Nemotron 3 Nano 30B A3B | Not in LiveBench 2026-06-25 | 1,314 | ~32 GBRuns slowly | 24.6 | NVIDIA Nemotron Open Model | 1.4M |
| Qwen3 30B A3B Instruct 2507 | Not in LiveBench 2026-06-25 | 1,382 | ~25 GBRuns slowly | 18.6 | Apache 2.0 | 1.1M |
| GLM-4.7 Flash | Not in LiveBench 2026-06-25 | 1,365 | ~25 GBRuns slowly | 18.3 | MIT | 953.7K |
| Nemotron 3.5 Lightning 30B A3B | Not in LiveBench 2026-06-25 | Arena rated the NVFP4 build, NVIDIA-Nemotron-3.5-Lightning-30B-A3B (NVFP4); this entry is the BF16 reference, which NVIDIA reports with different scores | ~25 GBRuns slowly | 18.9 | OpenMDW 1.1 | 861.5K |
| Qwen3 Coder 30B A3B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~25 GBRuns slowly | 18.6 | Apache 2.0 | 745K |
| Devstral Small 2 24B | Not in LiveBench 2026-06-25 | Not on the LMArena text leaderboard of 2026-10-22 | ~20 GBRuns slowly | 14.3 | Apache 2.0 | 195.6K |
| Mistral Small 3.2 24B | Not in LiveBench 2026-06-25 | 1,356 | ~21 GBRuns slowly | 14.3 | Apache 2.0 | 181.4K |
| Granite 4.2 30B | Not in LiveBench 2026-06-25 | 1,340 | ~25 GBRuns slowly | 17.7 | Apache 2.0 | 39.3K |
| Olmo 3 32B Think | Not in LiveBench 2026-06-25 | 1,306 | ~26 GBRuns slowly | 19.5 | Apache 2.0 | 13.1K |
| Smaug Mini | 76.9 | Not on the LMArena text leaderboard of 2026-10-22 | ~23 GBRuns slowly | 16.5 | Apache 2.0 | 817 |
Also runs slowly
Above the 16 GB of VRAM, the layers that do not fit run on the CPU from PC RAM, up to 40 GB in all. 32 GB of PC RAM assumed: this configuration lists none. Planning estimate, not a speed figure. The 16 reviewed models that run slowly are in the table above.
- txgemma-27b-chatgoogle~23 GBruns well with 24 GB
- txgemma-27b-predictgoogle~23 GBruns well with 24 GB
- granite-4.1-30bibm-granite~24 GBruns well with 24 GB
- granite-4.1-30b-baseibm-granite~24 GBruns well with 24 GB
- Qwen3-30B-A3B-Thinking-2507Qwen~25 GBQ4 · tensor count
- DeepSeek-R1-Distill-Qwen-32Bdeepseek-ai~27 GBQ4 · tensor count
- SWE-Dev-32Bzai-org~27 GBQ4 · tensor count
- SynLogic-32BMiniMaxAI~27 GBQ4 · tensor count
- SynLogic-Mix-3-32BMiniMaxAI~27 GBQ4 · tensor count
- DRIVE-SFTtencent~27 GBQ4 · tensor count
- DRIVE-RLtencent~27 GBQ4 · tensor count
- Smaug-34B-v0.1abacusai~28 GBQ4 · tensor count
- CodeLlama-34b-hfmeta-llama~28 GBQ4 · tensor count
- CodeLlama-34b-Instruct-hfmeta-llama~28 GBQ4 · tensor count
- MetaMath-Bagel-DPO-34Babacusai~28 GBQ4 · tensor count
- MetaMath-bagel-34b-v0.2-c1500abacusai~28 GBQ4 · tensor count
- MM-OV-bagel-DPO-34b-c1000-250abacusai~28 GBQ4 · tensor count
- MM-Orc-Vic-bagel-34b-c1000abacusai~28 GBQ4 · tensor count
- CodeLlama-34b-Python-hfmeta-llama~28 GBQ4 · tensor count
- Smaug-Mixtral-v0.1abacusai~38 GBQ4 · tensor count
Adjust the estimate: quantization, context, reserve
Vision
Vision · LMArena
rating · higher is better
- Color = publisher
Source: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0 · Models without a rating are left out, not scored as zero.
| Molmo2-8BQ4 · Hugging Face count · source | 1,112 ± 20 | ~9 GB | Q4 · Hugging Face count |
|---|---|---|---|
| Qwen2-VL-7B-InstructQ4 · Hugging Face count · source | 1,031 ± 11 | ~9 GB | Q4 · Hugging Face count |
| Molmo-7B-D-0924Q4 · Hugging Face count · source | 994 ± 14 | ~8 GB | Q4 · Hugging Face count |
| Llama-3.2-11B-Vision-InstructQ4 · Hugging Face count · source | 993 ± 11 | ~11 GB | Q4 · Hugging Face count |
| MiniCPM-V-2_6Q4 · Hugging Face count · source | 962 ± 16 | ~9 GB | Q4 · Hugging Face count |
| Phi-3.5-vision-instructQ4 · Hugging Face count · source | 919 ± 16 | ~6 GB | Q4 · Hugging Face count |
| Gemma 4 12BQ4 · published params · source | Not on the LMArena vision leaderboard of 2026-10-19 | ~11 GB | Q4 · published params |
| Qwen3.5 9BQ4 · published params · source | Not on the LMArena vision leaderboard of 2026-10-19 | ~9 GB | Q4 · published params |
LMArena ratings in this table: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0
Also runs slowly
Above the 16 GB of VRAM, the layers that do not fit run on the CPU from PC RAM, up to 40 GB in all. 32 GB of PC RAM assumed: this configuration lists none. Planning estimate, not a speed figure.
- Gemma 4 26B A4BGoogle~21 GBruns well with 24 GB
- Qwen3.6 27BAlibaba~23 GBruns well with 24 GB
- Qwen3.8 27BAlibaba~23 GBruns well with 24 GB
- Smaug MiniAbacus AI~23 GBruns well with 24 GB
- Gemma 4 31BGoogle~26 GBQ4 · published params
- Qwen3.5 35B A3BAlibaba~29 GBQ4 · published params
- diffusiongemma-26B-A4B-itgoogle~22 GBruns well with 24 GB
- deepseek-vl2deepseek-ai~23 GBruns well with 24 GB
- Fara1.5-27Bmicrosoft~23 GBruns well with 24 GB
- gemma-3-27b-itgoogle~23 GBruns well with 24 GB
- ERNIE-4.5-VL-28B-A3B-PTbaidu~25 GBQ4 · Hugging Face count
- ERNIE-4.5-VL-28B-A3B-Thinkingbaidu~25 GBQ4 · Hugging Face count
- Keye-VL-2.0-30B-A3BKwai-Keye~26 GBQ4 · Hugging Face count
- Qwen3-VL-30B-A3B-InstructQwen~26 GBQ4 · Hugging Face count
- gemma-4-31Bgoogle~27 GBQ4 · Hugging Face count
- Qwen2.5-VL-32B-InstructQwen~28 GBQ4 · Hugging Face count
- Qwen3.6-35B-A3BQwen~29 GBQ4 · Hugging Face count
12 more vision models in the catalog have no usable memory figure or parameter count for this profile. Browse Vision models
Image
Image · Prompt alignment · GenEval
0–1 · higher is better
Source: GenEval · 2023 · reviewed 2026-10-05 · Published quality in the original benchmark preset; not measured on this hardware.
| Stable Diffusion XL 1.0Platform path documented · stable-diffusion.cpp · source | 0.55 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|---|
| Stable Diffusion 1.5Platform path documented · stable-diffusion.cpp · source | 0.43 | Not on the LMArena text-to-image leaderboard of 2026-10-07 | Documented path without a memory figure | Platform path documented |
| TRELLIS-image-largePublisher minimum · CUDA · source | No comparable published result | Not reviewed for LMArena | ~16 GB | Publisher minimum |
| FLUX.2 [klein] 4BPublisher minimum · CUDA · source | No comparable published result | 1,030 ± 3 | ~13 GB | Publisher minimum |
| FLUX.2 [klein] Base 4BPublisher minimum · CUDA · source | No comparable published result | Arena rated flux-2-klein-4b; this entry is the undistilled Base checkpoint | ~13 GB | Publisher minimum |
| BagelConditional path · CUDA · source | No comparable published result | 898 ± 6 | ~12 GB | Conditional path |
| Stable Diffusion 3.5 LargeConditional path · CUDA · source | No comparable published result | 938 ± 5 | ~11 GB | Conditional path |
| Stable Diffusion 3.5 MediumConditional path · source | No comparable published result | Not on the LMArena text-to-image leaderboard of 2026-10-07 | ~10 GB | Conditional path |
LMArena ratings in this table: LMArena 2026-10-07 · collected 2026-10-25 · dataset · CC BY 4.0
With 24 GB you also run… 23
The next memory size on this ladder. Figures are the requirement on this hardware type.
- Kandinsky-5.0-T2I-Lite-sft-Diffuserskandinskylab~17 GB16-bit est. · params
- LongCat ImageMeituan~17 GBPublisher minimum
- Z-ImageTongyi-MAI~17 GB16-bit est. · params
- Janus ProDeepSeek~19 GB16-bit est. · params
- Qwen-Image-2.1Qwen~20 GB16-bit est. · params
- Qwen-Image-2.1-TurboQwen~20 GB16-bit est. · params
- FIBOBria~22 GB16-bit est. · params
- FIBO LiteBria~22 GB16-bit est. · params
- Fibo-1.5briaai~22 GB16-bit est. · params
- Fibo-Editbriaai~22 GB16-bit est. · params
- HiDream-O1-ImageHiDream~22 GB16-bit est. · params
- HiDream-O1-Image-DevHiDream~22 GB16-bit est. · params
- Infinity 8BByteDance~22 GB16-bit est. · params
- Stable Diffusion 3.5 Large TurboStability AI~22 GB16-bit est. · params
- GLM-ImageZ AI~23 GBPublisher minimum
- ERNIE ImageBaidu~24 GBPublisher minimum
- ERNIE Image TurboBaidu~24 GBPublisher minimum
- FLUX.2 [dev]Black Forest Labs~24 GBPublisher minimum
- HiDream-O1-Image-Dev-2604HiDream-ai~24 GB16-bit est. · params
- HunyuanImage-2.1Tencent Hunyuan~24 GBPublisher minimum
- Ideogram 4.0Ideogram~24 GBPublisher minimum
- Ideogram 4.0 (Quality)Ideogram~24 GBPublisher minimum
- TRELLIS.2-4Bmicrosoft~24 GBPublisher minimum
17 more image models in the catalog have no usable memory figure or parameter count for this profile. Browse Image models
Video
| CogVideoX-5b-I2V16-bit est. · params · source | ~16 GB | 16-bit est. · params |
|---|---|---|
| CogVideoX1.5-5B16-bit est. · params · source | ~16 GB | 16-bit est. · params |
| CogVideoX1.5-5B-I2V16-bit est. · params · source | ~16 GB | 16-bit est. · params |
| SANA-Video_2.0_5B_720p16-bit est. · params · source | ~13 GB | 16-bit est. · params |
| Kandinsky-6.0-Lite-distill-5s-Diffusers16-bit est. · params · source | ~10 GB | 16-bit est. · params |
| Allegro16-bit est. · params · source | ~9 GB | 16-bit est. · params |
| Wan2.1 T2V 1.3BPlatform path documented · Wan2.1 / PyTorch · source | ~8 GB | Platform path documented |
| Wan2.1-T2V-1.3B-DiffusersConditional path · CUDA · source | ~8 GB | Conditional path |
How this list works Documented execution paths on this hardware first, then publisher minimums (CUDA or Apple MPS).
With 24 GB you also run… 3
The next memory size on this ladder. Figures are the requirement on this hardware type.
- HunyuanVideo-1.5tencent~22 GB16-bit est. · params
- Wan2.2-TI2V-5BWan-AI~24 GBPublisher minimum
- Wan2.2-TI2V-5B-DiffusersWan-AI~24 GBPublisher minimum
20 more video models in the catalog have no usable memory figure or parameter count for this profile. Browse Video models
Audio
Audio · Word error rate · Open ASR
% · lower is better
Source: Word error rate · Open ASR · reviewed 2026-10-20 · Published quality in the original benchmark preset; not measured on this hardware.
| Whisper large-v3-turboPlatform path documented · whisper.cpp / CUDA · source | 7.83 | Documented path without a memory figure | Platform path documented |
|---|---|---|---|
| Whisper large-v3Platform path documented · whisper.cpp / CUDA · source | 7.44 | Documented path without a memory figure | Platform path documented |
| GaussianMind16-bit est. · params · source | No comparable published result | ~16 GB | 16-bit est. · params |
| Higgs Audio V3 TTSPublisher minimum · CUDA · source | No comparable published result | ~16 GB | Publisher minimum |
| higgs-tts-2-3b-base16-bit est. · params · source | No comparable published result | ~16 GB | 16-bit est. · params |
| Maya1Publisher minimum · CUDA · source | No comparable published result | ~16 GB | Publisher minimum |
| MusicGen SmallPlatform path documented · AudioCraft / PyTorch · source | No comparable published result | ~16 GB | Platform path documented |
| Phi-4-multimodal-instruct16-bit est. · params · source | No comparable published result | ~16 GB | 16-bit est. · params |
With 24 GB you also run… 12
The next memory size on this ladder. Figures are the requirement on this hardware type.
- omniASR-W2V-7Bfacebook~18 GB16-bit est. · params
- MiMo-V2.5-ASRXiaomiMiMo~21 GB16-bit est. · params
- moshiko-pytorch-bf16kyutai~21 GB16-bit est. · params
- audio-flamingo-3-hfnvidia~22 GB16-bit est. · params
- Llasa-8BHKUSTAudio~23 GB16-bit est. · params
- MOSS-TTS-v1.5OpenMOSS-Team~23 GB16-bit est. · params
- personaplex-7b-v1nvidia~23 GB16-bit est. · params
- Qwen2-Audio-7B-InstructQwen~23 GB16-bit est. · params
- VibeVoice-ASRmicrosoft~23 GB16-bit est. · params
- VibeVoice-ASR-Streaming-7Bmicrosoft~23 GB16-bit est. · params
- Fish Audio S2 ProFish Audio~24 GBPublisher minimum
- MiniMax-Music3MiniMaxAI~24 GBPublisher minimum
33 more audio models in the catalog have no usable memory figure or parameter count for this profile. Browse Audio models
Embeddings
Embeddings · MTEB (Multilingual, v2) · mean(task)
0–100 · higher is better
- Color = publisher
| Qwen3-Embedding 8B8-bit · published params · source | 70.6 | ~12 GB | 8-bit · published params |
|---|---|---|---|
| Qwen3-Embedding 4B8-bit · published params · source | 69.5 | ~7 GB | 8-bit · published params |
| Harrier OSS v1 (0.6B)8-bit · published params · source | 69.0 | ~3 GB | 8-bit · published params |
| Qwen3-Embedding 0.6B8-bit · published params · source | 64.3 | ~3 GB | 8-bit · published params |
| Multilingual E5 Large Instruct8-bit · published params · source | 63.2 | ~3 GB | 8-bit · published params |
| EmbeddingGemma 28-bit · published params · source | 61.4 | ~3 GB | 8-bit · published params |
| EmbeddingGemma (300M)8-bit · published params · source | 61.1 | ~3 GB | 8-bit · published params |
| BGE-M38-bit · published params · source | 59.6 | ~3 GB | 8-bit · published params |
With 24 GB you also run… 1
The next memory size on this ladder runs these well. Figures are the requirement on this hardware type.
- KaLM-Embedding-Gemma3-12B-2511tencent~17 GB8-bit · Hub count
Runs well within 16 GB of VRAM · planning estimates, not measured speed
Measurements
Published tokens per second, exactly as measured; never estimated.
| gpt-oss-20b MXFP4 | 186.51 ± 0.33 | MXFP4 (native gpt-oss weights) | llama.cpp CUDA | llama.cpp gpt-oss guide · maintainer result |
|---|---|---|---|---|
| Llama 2 7B Q4_0 | 150.20 ± 0.26 | Q4_0 | Vulkan | llama.cpp Vulkan scoreboard · community result |
| Llama 2 7B Q4_0 | 147.48 ± 1.41 | Q4_0 | CUDA | llama.cpp CUDA scoreboard · community result |
Specifications
20 sourced fields; the source is named once per group, and a field from another source links to it. Official specifications
Marker: the lightest text model that does not run well here but runs well with 24 GB. No text model that runs well here has a LiveBench score. Q4 planning estimate, not a measured fit.
Reviewed · Manufacturer source
- Graphics processor
- NVIDIA GeForce RTX 4080 SUPER
- GPU runtime status
- Not stated by the source
- Memory capacity
- 16 GB
- Memory kind
- Dedicated VRAM
- Power connectors
- 3x PCIe 8-pin cables (adapter in box) OR / 1x 450 W or greater PCIe Gen 5 cable
- AI performance (TOPS, manufacturer)
- 836 TOPS
- GPU architecture
- Ada Lovelace
- GPU base clock
- 2,290 MHz
- GPU boost clock
- 2,550 MHz
- CUDA compute capability
- 8.9
- CUDA cores
- 10,240
- GPU memory type
- GDDR6X
- GPU power
- 320 W
- Card length
- 304 mm
- Memory interface
- 256 bit
- Minimum power supply
- 750 W
- PCIe interface
- Gen 4
- Slot width
- 3
NVIDIA GeForce RTX 4080 SUPER with 16 GB VRAM. Runtime support and local inference speed are not verified.
Photo: NVIDIA family photo · memory configuration is not pictured. Founders Edition reference design; partner boards vary. · Manufacturer source
Memory bandwidth is not measured model speed; verify runtime and checkpoint compatibility. Hardware specification reviewed: 2026-10-13.
How estimates work
- Candidates
- Reviewed pages use published parameter counts; automatic text candidates use official repository tensor counts (packed quantized weights and unsupported architectures excluded).
- Planning
- Q4 weights + 20% allowance + reserve: 25% of unified or system memory (at least 2 GB, at most 8 GB; 2 GB here) or 2 GB per dedicated GPU. Adjustable in “Adjust the estimate”. No KV cache is calculated from context.
- Fit tiers
- Runs well: the estimate fits the fast memory (VRAM, or unified memory within the macOS GPU share). Runs slowly: within the machine's memory but outside the fast part; a graphics card offloads layers to PC RAM (the listed RAM, or 32 GB assumed). Doesn't run: above both. Tiers are planning estimates, never a speed.
- Vision
- Vision-language models use the Q4 planning estimate from the parameter count (published by the model page, or read from Hugging Face). Classic vision models (classify, detect, segment) use a 16-bit estimate. No vision benchmark is tracked, so rows carry no score.
- Media and embeddings
- Media rows use documented execution paths (exact hardware or platform) and publisher minimums; a published requirement above this memory moves a model to the next tier. Without either, a 16-bit estimate from the published parameter count (× 1.2 + 2 GB), labelled as such. Embeddings use 8-bit weights + 20% + 2 GB.
Local AI basics: parameters, Q4 and memory · Full model catalog · Data and methodology
Alternatives
Side-by-side comparisons, then other machines on the same memory ladder (VRAM); each page lists the models that fit it.
Compare with… 1
Also 16 GB 8
Less memory · 12 GB 8
- RTX 5070 · 12 GB VRAM
- RTX 3060 · 12 GB VRAM
- RTX 4070 · 12 GB VRAM
- RTX 4070 SUPER · 12 GB VRAM
- Dell Alienware 16X Aurora (AC16251) · Core Ultra 9 290HX Plus · RTX 5070 Ti 12 GB · 32 GB
- HP ZBook Fury 16 G11 · Core i9-14900HX · RTX 4000 Ada 12 GB · 32 GB
- ROG Zephyrus G14 (2025) · Ryzen AI 9 HX 370 · RTX 5070 Ti · 32 GB
- RTX 5070 Laptop GPU · 12 GB VRAM
