Gemma 4 12B
Local OKExcellentA · 44
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: est. 71 tok/s, typical error 15% · first token ~0.3s
Fits in 16 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 4080.
Best Local AI Models for RTX 4080 (16GB)
The RTX 4080 has 16GB of GDDR6X VRAM at 716.8 GB/s. That fits dense models up to 14B at Q4, with an 8B model at est. 100 tok/s, typical error 5% and a 14B model at est. 62 tok/s, typical error 15%. The RTX 4080 SUPER that followed it keeps the 16GB and raises bandwidth to 736 GB/s, so the two run the same models at nearly the same speed.
Quick answer
The best local LLM for the RTX 4080 is Gemma 4 12B (est. 71 tok/s, typical error 15%) on its 16GB VRAM. It uses ~8GB of VRAM; the RTX 4080 handles dense models up to ~14B at Q4. A 14B model at Q4 runs at est. 62 tok/s, typical error 15%. (ModelFit, 2026)
Sizing rule: a Q4 model needs about 0.6 GB of VRAM per billion parameters, and ModelFit budgets 90% of this card's 16GB for weights, context, and KV-cache. The per-size table below uses that same budget. Other strong fits: Qwen3.5 9B Instruct (Q8) (9B, ~10.7GB) and Mellum2.1 12B-A2.5B Thinking (12.1B, ~8.1GB). On this card Gemma 4 12B runs at est. 71 tok/s, typical error 15%. What it will not run: a 32B model does not fit in VRAM alone, so it spills to system RAM over PCIe and slows sharply.
$ollama run gemma4:12b
- Top pick
- Gemma 4 12B
- Decode speed
- est. 71 tok/s, typical error 15%
- VRAM needed
- ~8 GB
Speeds are ModelFit estimates from memory bandwidth and model size, not measured benchmarks. Typical error: the median gap between the ModelFit speed model and measured runs on hardware the model was not fitted on, for this backend and model size.
Cite this page: ModelFit, RTX 4080 VRAM: 16GB GDDR6X for Local LLMs, 14B at est. 62 tok/s, https://modelfit.io/gpu/rtx-4080/, updated October 2026, CC BY 4.0.
Specs
VRAM
16 GB GDDR6X
Speed (8B Q4)
est. 100 tok/stypical error 5%
Bandwidth
716.8 GB/s
Architecture
Ada Lovelace
Max model size
Up to ~14B dense
Compatibility
10 excellent, 0 workable
Vendor
Table 1
RTX 4080 estimated tokens/sec by model size
Q4_K_M · ModelFit estimate| Model size | Est. speed | Fit on 16GB |
|---|---|---|
| 7B | est. 114 tok/stypical error 5% | Fits in VRAM |
| 14B | est. 62 tok/stypical error 15% | Fits in VRAM |
| 20B MoE (3.6B active) | est. 131 tok/stypical error 15% | Fits in VRAM |
| 32B | est. 6 tok/srough estimate, offload not measured | CPU offload (slow) |
| 35B MoE (3B active) | est. 53 tok/srough estimate, offload not measured | CPU offload (slow) |
| 70B | est. 1 tok/srough estimate, offload not measured | CPU offload (slow) |
| 120B MoE (5.1B active) | est. 34 tok/srough estimate, offload not measured | CPU offload (slow) |
Fig. 1RTX 4080: estimated speed by model size
Decode speed, generic Q4_K_M model of each size at 4K context, 716.8 GB/s, typical error 5-15%
Fits in 16 GBCPU offload (slow)
7B114
14B62
20B MoE (3.6B active)131
32B6
35B MoE (3B active)53
70B1
120B MoE (5.1B active)34
tok/s est.ModelFit estimates, not measured benchmarks: decode speed of a generic Q4_K_M model of each size at 4K context on the RTX 4080's 716.8 GB/s bandwidth. MoE rows scale by active parameters (decode reads only the active experts), so a 35B MoE runs far faster than a dense 32B. "CPU offload" sizes exceed the 16GB VRAM; dense models slow to a crawl there, MoE models degrade less because hot experts stay GPU-resident. Offload rows are rough estimates: no measured run covers offload, so they carry no typical error. What typical error means.
Fig. 2
Context costs VRAM too.
Gemma 4 12B loads ~8 GB of weights; at 16k context the KV cache adds ~3.0 GB (still fits the ~14 GB usable VRAM), and at 64k it adds ~12.0 GB (exceeds the budget, use a smaller quant or a q8_0 KV cache).
Weights 8.0 GBKV cache · 16k context 3.0 GBHeadroom 3.4 GBRuntime reserve 1.6 GB
The model leaves about 3.4 GB of the usable VRAM budget free after weights and 16k context.
KV-cache figures assume an fp16 cache, the llama.cpp/Ollama default. Standard GQA models use a size-class estimate (8 KV heads x 128 head dim class); hybrid linear-attention models (Qwen3.5/3.6, Qwen3-Next) use the exact per-token cost from their published config, since only their sparse full-attention layers cache KV. A q8_0 KV cache roughly halves either figure. Estimates, not measurements.
Get told when a model beats Gemma 4 12B on your RTX 4080
One email when a new open-weight model is a better fit for 16 GB of VRAM, plus the Thursday weekly.
You'll also get the weekly ModelFit email. By signing up you agree to our Privacy Policy. Unsubscribe from either email anytime, in one click.
Where to buy the RTX 4080
Storage & accessories for your model library
- Internal NVMe SSD · 2TBA Gen4 M.2 drive keeps your whole GGUF and quant collection on fast local storage, loading models straight off NVMe.Check price on Amazon
- USB4 NVMe Enclosure40Gbps external storage fast enough to run models from. Pair it with an M.2 drive for a portable model vault.Check price on Amazon
ModelFit may earn a commission on purchases through these links, at no extra cost to you.
RTX 4080 VRAM for AI: what actually fits?
ModelFit budgets about 14.4GB of the 16GB for weights and context. 14B models at Q4 (~11GB) load with room for a long context, and 20B-class mixture-of-experts models fit too. Dense 27B and 32B models do not: they spill into system RAM. Against the RTX 5080 (960 GB/s, est. 132 tok/s, typical error 5% on 8B) the 4080 gives up speed but not capacity, since both are 16GB cards. For 27B-32B models fully in VRAM, a 24GB card such as the RTX 3090 is the step up.
Table 2
RTX 4080 vs similar GPUs
| Hardware | Memory | Speed | Bandwidth |
|---|---|---|---|
| RTX 4070 Ti SUPER | 16 GB | est. 94 tok/stypical error 5% | 672 GB/s |
| RTX 4080 | 16 GB | est. 100 tok/stypical error 5% | 716.8 GB/s |
| RTX 4080 SUPER | 16 GB | est. 102 tok/stypical error 5% | 736 GB/s |
| RTX 5080 | 16 GB | est. 132 tok/stypical error 5% | 960 GB/s |
| RTX 3090 | 24 GB | est. 117 tok/stypical error 5% | 936 GB/s |
Recommended models
10 modelsQwen3.5 9B Instruct (Q8)
Local OKOKB · 26
Best for: Quality, Coding, Reasoning·Pop: 86/100
Perf: est. 68 tok/s, typical error 5% · first token ~1.1s
Fits in 16 GB VRAM with room to spare. Best for quality, coding, reasoning on RTX 4080.
Mellum2.1 12B-A2.5B Thinking
Local OKExcellentA · 44
Best for: Agentic coding, Small machines·Pop: 76/100
Perf: est. 186 tok/s, typical error 15% · first token ~0.1s
Fits in 16 GB VRAM with room to spare. Best for agentic coding, small machines on RTX 4080.
Gemma 4 E4B (Q8)
Local OKExcellentA · 48
Best for: On-device, Mobile, Chat·Pop: 82/100
Perf: est. 107 tok/s, typical error 5% · first token ~0.9s
Fits in 16 GB VRAM with room to spare. Best for on-device, mobile, chat on RTX 4080.
Gemma 4 12B (Q8)
Local OKOKB · 11
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: est. 45 tok/s, typical error 15% · first token ~1.5s
Fits in 16 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 4080.
NVIDIA Nemotron 3.5 Bolt E4B
Local OKExcellentA · 45
Best for: Voice assistants, Tool calling, Chat·Pop: 75/100
Perf: est. 132 tok/s, typical error 5% · first token ~0.2s
Fits in 16 GB VRAM with room to spare. Best for voice assistants, tool calling, chat on RTX 4080.
Ministral 3 14B
Local OKExcellentA · 37
Best for: Chat, Reasoning, Vision·Pop: 74/100
Perf: est. 62 tok/s, typical error 15% · first token ~0.3s
Fits in 16 GB VRAM with room to spare. Best for chat, reasoning, vision on RTX 4080.
Qwen3.5 9B Instruct
Local OKExcellentA · 51
Best for: Quality, Coding, Reasoning·Pop: 86/100
Perf: est. 106 tok/s, typical error 5% · first token ~0.2s
Fits in 16 GB VRAM with room to spare. Best for quality, coding, reasoning on RTX 4080.
LFM2 24B-A2B Instruct
Local OKOKB · 3
Best for: Local AI agents, privacy-first tool calling, MCP workflows·Pop: 80/100
Perf: est. 203 tok/s, typical error 15% · first token ~0.1s
Fits in 16 GB VRAM with room to spare. Best for local ai agents, privacy-first tool calling, mcp workflows on RTX 4080.
Ornith 1.0 9B
Local OKExcellentA · 61
Best for: Agentic coding on small machines·Pop: 76/100
Perf: est. 106 tok/s, typical error 5% · first token ~0.2s
Fits in 16 GB VRAM with room to spare. Best for agentic coding on small machines on RTX 4080.
Video and image models on the RTX 4080
Runs · Stated: the linked source says a GPU with that much VRAM runs the model. Runs · Published: a dated run on the named card with that much VRAM. unknown: no source yet. Exceeds VRAM: even the main weights are too large for 16 GB. Every VRAM tier: video · image. Ranked with reasons: video models for 16 GB of VRAM · image models for 16 GB of VRAM.
Video
| Model | Fits | Quality | App |
|---|---|---|---|
| HunyuanVideo-1.5Tencent Hunyuan Community License | Runs · Stated14 GB VRAM · official runner with model offloading · HunyuanVideo-1.5 model card | #38 of 48 in text to video4,292 votessnapshot 2026-09-22 | ComfyUI |
| LTX-2.3LTX-2 Community License Agreement | Runs · Stated16 GB VRAM · distilled with FP8 cast, previews only · LTX blog: hardware for AI video models | No independent score yet | Lightricks/LTX-2 ltx-pipelines, ComfyUI |
| LTX-2.5LTX-2.x Community License Agreement | Runs · Stated12 GB VRAM · LTX Desktop on Windows, weight streaming, 40 GB RAM, 540p · LTX Desktop compatibility matrix (Lightricks) | No independent score yet | Lightricks/LTX-2 ltx-pipelines, ComfyUI |
| MiniMax H3MiniMax H3 Community License Agreement | Runs · PublishedRTX 5060 · 16 GB VRAM · ComfyUI int8 convrot with int4 text encoder, 1024x768, 10 s in 24 min · ComfyUI issue #15274 | #8 of 48 in text to video11,417 votessnapshot 2026-09-22 | stable-diffusion.cpp, ComfyUI |
| Wan2.1Apache License 2.0 | Runs · Stated13 GB VRAM · T2V 14B in Diffusers, bf16 with group offloading to system RAM · Diffusers docs: Wan pipeline | No independent score yet | ComfyUI |
| Wan2.2 A14BApache License 2.0 | Runs · PublishedRTX 5060 Ti · 16 GB VRAM · ComfyUI fp16 I2V with 4-step LoRAs, 64 GB system RAM · ComfyUI issue #10662 | #42 of 48 in text to video10,458 votessnapshot 2026-09-22 | ComfyUI |
| Wan2.2 TI2V 5BApache License 2.0 | Runs · Stated8 GB VRAM · ComfyUI native offloading · ComfyUI docs: Wan2.2Complete: Comfy-Org repackaged FP16 with FP8 scaled UMT518.1 GB download | No independent score yet | ComfyUI |
| Mochi 1 previewApache License 2.0 | unknownComplete: calcuis GGUF Q3_K_M with Q4_0 T5 and FP8 VAE7.7 GB download | #48 of 48 in text to video5,893 votessnapshot 2026-09-22 | Diffusers, ComfyUI |
Quality: blind human votes from Arena (arena.ai), CC BY 4.0, snapshot 2026-09-22. Scores are for the Arena checkpoint, not your local quantized build.
Image
| Model | Fits | Quality | App |
|---|---|---|---|
| FLUX 3 Image KleinFLUX Kommunity License v1.0 | Runs · Stated8 GB VRAM · official ComfyUI int8 files · FLUX 3 Image Klein README (Black Forest Labs) | No independent score yet | Diffusers |
| FLUX.1 [dev]FLUX.1 [dev] Non-Commercial License v1.1.1 | Runs · Stated6 GB VRAM · stable-diffusion.cpp with GGUF weights · stable-diffusion.cpp FLUX docs | #78 of 82 in text to image49,221 votessnapshot 2026-10-07 | Diffusers, ComfyUI |
| FLUX.2 [klein]Apache 2.0 (4B), FLUX Non-Commercial License (9B) | Runs · Stated13 GB VRAM · Klein 4B · FLUX.2 [klein] 4B model cardComplete: Klein 4B BF1616.0 GB download | No independent score yet | Diffusers, stable-diffusion.cpp |
| Krea 2 TurboKrea 2 Community License Agreement v.1 | Runs · PublishedRTX 4070 · 12 GB VRAM · ComfyUI fp8, dynamic VRAM paging, 8 steps · ComfyUI issue #15365 | No independent score yet | Diffusers, stable-diffusion.cpp |
| Qwen-Image-2.1Qwen Research License Agreement | Runs · PublishedRTX 5080 · 16 GB VRAM · stable-diffusion.cpp Q8_0, 1024x1024 · stable-diffusion.cpp issue #2041 | #19 of 82 in text to image11,441 votessnapshot 2026-10-07 | Diffusers, stable-diffusion.cpp |
| SDXL Base 1.0CreativeML Open RAIL++-M License | Runs · Stated8 GB VRAM · SDXL 1.0, no app named · SDXL 1.0 announcement (Stability AI) | No independent score yet | ComfyUI, stable-diffusion.cpp |
| Stable Diffusion 3.5 LargeStability AI Community License | Runs · Stated11 GB VRAM · FP8 TensorRT build on RTX 40 and 50 · Stability AI: SD3.5 TensorRT models | #81 of 82 in text to image23,396 votessnapshot 2026-10-07 | ComfyUI |
| Z-Image-TurboApache 2.0 | Runs · Stated16 GB VRAM · stated for consumer GPUs · Z-Image-Turbo model card | #59 of 82 in text to image24,808 votessnapshot 2026-10-07 | Diffusers, stable-diffusion.cpp |
| MOSS-ImageApache 2.0 | unknownComplete: UD-Q4_K_XL GGUF13.2 GB download | No independent score yet | stable-diffusion.cpp |
| FLUX.2 [dev]FLUX Non-Commercial License | Exceeds VRAMComplete: Q4_K_M GGUF34.8 GB download | #38 of 82 in text to image77,988 votessnapshot 2026-10-07 | Diffusers, ComfyUI, stable-diffusion.cpp |
Quality: blind human votes from Arena (arena.ai), CC BY 4.0, snapshot 2026-10-07. Scores are for the Arena checkpoint, not your local quantized build.
Similar GPUs for local AI
- RTX 4080 SUPER16GB · est. 102 tok/s
- RTX 508016GB · est. 132 tok/s
- RTX 4070 Ti SUPER16GB · est. 94 tok/s
- RTX 309024GB · est. 117 tok/s
Speeds are estimates for Qwen3 8B Q4_K_M, typical error 5%.
Compatible model families
- QwenAlibaba Cloud: Qwen3.8 27B local flagship (Apache 2.0, vision)
- LlamaMeta: Most popular open-weight model family
- DeepSeekDeepSeek AI: V4.1 Flash: multimodal 552B MoE, API only
- MistralMistral AI: Excellent performance-per-parameter ratio
- GemmaGoogle DeepMind: Gemma 4 E4B: 4.5B effective, runs on 8GB Macs
- PhiMicrosoft: Best quality-per-gigabyte at small sizes
RTX 4080 FAQ: common questions
How much VRAM does the RTX 4080 have?
The RTX 4080 has 16GB of GDDR6X VRAM at 716.8 GB/s, the same capacity as the RTX 4080 SUPER. ModelFit budgets about 14.4GB of it for weights and context, enough for dense models up to 14B at Q4 quantization.
What size LLM can I run on an RTX 4080?
Dense models up to 14B at Q4, such as Qwen3 14B (~11GB) at est. 61 tok/s, typical error 15%, plus 20B-class mixture-of-experts models. 8B models run at est. 100 tok/s, typical error 5%. Dense 27B and 32B models need a 24GB card to stay in VRAM.
RTX 4080 vs RTX 4080 SUPER for local LLMs?
Nearly the same card for local AI: both have 16GB, and the SUPER streams 736 GB/s against 716.8 GB/s, so it is 2% faster on 8B models (est. 102 tok/s, typical error 5% vs est. 100 tok/s). Buy whichever is cheaper.
RTX 4080 vs RTX 3090 for local AI?
The RTX 3090 has 24GB against 16GB, so it holds 27B and 32B models the 4080 cannot. It also streams more memory (936 vs 716.8 GB/s): est. 117 tok/s, typical error 5% against est. 100 tok/s on 8B. The 4080 is the newer, more efficient card; for model size, the 3090 wins.
How fast is a 27B-class model on the RTX 4080?
The RTX 4080's 16GB of VRAM cannot fit a 32B model comfortably. The largest size class it fits is 20B MoE (3.6B active), at est. 131 tok/s, typical error 15%. The largest individual model in the catalog that fits is LFM2 24B-A2B Instruct (24B).
Related guides & benchmarks
More local AI guides
- Best LLM for 16GB VRAMEvery 16 GB card side by side, each with its best pick and speed.
- How Much RAM (or VRAM) Do You Need for a Local LLM?The model-size-to-memory matrix: what each VRAM and RAM tier actually runs.
- Best LLM for MacBook (Apple Silicon)Unified-memory Macs run bigger models per dollar. Picks by RAM tier, M1 to M5.
- Local AI image generator on MacCompare image models by RAM and app.
- Local image generation on MacImage models and apps by Mac RAM.
Browse all NVIDIA GPUs for AI
Real-world calibration
Have this hardware? Share a measured run.
Time a model that fits your machine with Ollama, using the same prompts as every other run. The CLI shows your result and asks before it sends anything; your machine then joins the speed test leaderboard. It does not change the quality benchmark.
$ npx @modelfit/cli@latest bench
Want personalized recommendations?
Use our interactive wizard to compare models across Apple Silicon and NVIDIA GPUs.