Gemma 4 12B
Local OKExcellentA · 44
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: est. 96 tok/s, typical error 15% · first token ~0.2s
Fits in 16 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 5080.
Best Local AI Models for RTX 5080 (16GB)
The RTX 5080 is the fastest 16GB card for local AI. On 8B models it reaches est. 132 tok/s, typical error 5%, outperforming even the RTX 3090 in raw speed while costing less. The best choice for users who want top speed with 14B models.
Quick answer
The best local LLM for the RTX 5080 is Gemma 4 12B (est. 96 tok/s, typical error 15%) on its 16GB VRAM. It uses ~8GB of VRAM; the RTX 5080 handles up to 14B parameter models at Q4. A 14B model at Q4 runs at est. 83 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. 96 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. 96 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 5080 16GB Local LLM: Fastest 16GB, 14B at est. 83 tok/s, https://modelfit.io/gpu/rtx-5080/, updated October 2026, CC BY 4.0.
Specs
VRAM
16 GB GDDR7
Speed (8B Q4)
est. 132 tok/stypical error 5%
Bandwidth
960 GB/s
Architecture
Blackwell
Max model size
Up to 14B parameter models
Compatibility
10 excellent, 0 workable
Vendor
Table 1
RTX 5080 estimated tokens/sec by model size
Q4_K_M · ModelFit estimate| Model size | Est. speed | Fit on 16GB |
|---|---|---|
| 7B | est. 149 tok/stypical error 5% | Fits in VRAM |
| 14B | est. 83 tok/stypical error 15% | Fits in VRAM |
| 20B MoE (3.6B active) | est. 171 tok/stypical error 15% | Fits in VRAM |
| 32B | est. 8 tok/stypical error 5% | CPU offload (slow) |
| 35B MoE (3B active) | est. 69 tok/stypical error 15% | CPU offload (slow) |
| 70B | est. 2 tok/stypical error 10% | CPU offload (slow) |
| 120B MoE (5.1B active) | est. 45 tok/stypical error 15% | CPU offload (slow) |
Fig. 1RTX 5080: estimated speed by model size
Decode speed, generic Q4_K_M model of each size at 4K context, 960 GB/s, typical error 5-15%
Fits in 16 GBCPU offload (slow)
7B149
14B83
20B MoE (3.6B active)171
32B8
35B MoE (3B active)69
70B2
120B MoE (5.1B active)45
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 5080's 960 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. 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 5080
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 5080
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 5080 VRAM for AI: what actually fits?
16GB GDDR7 at 960 GB/s makes the RTX 5080 the bandwidth champion of 16GB cards. It loads 14B models at Q4 with ~5GB headroom and runs them at est. 83 tok/s, typical error 15%. On 8B models it pushes est. 132 tok/s, typical error 5%; for context, the RTX 3090 achieves est. 117 tok/s with 24GB VRAM. The 5080 is faster despite 8GB less memory. If your models fit in 16GB, this card maximizes speed. For models that need 20GB+, you will need to step up to the RTX 4090 or 5090.
Table 2
RTX 5080 vs similar GPUs
| Hardware | Memory | Speed | Bandwidth |
|---|---|---|---|
| RTX 4080 SUPER | 16 GB | est. 102 tok/stypical error 5% | 736 GB/s |
| RTX 5070 Ti | 16 GB | est. 125 tok/stypical error 5% | 896 GB/s |
| RTX 5080 | 16 GB | est. 132 tok/stypical error 5% | 960 GB/s |
| RTX 5090 | 32 GB | est. 213 tok/stypical error 5% | 1792 GB/s |
Recommended models
registry-verified10 modelsQwen3.5 9B Instruct (Q8)
Local OKOKB · 26
Best for: Quality, Coding, Reasoning·Pop: 86/100
Perf: est. 91 tok/s, typical error 5% · first token ~0.9s
Fits in 16 GB VRAM with room to spare. Best for quality, coding, reasoning on RTX 5080.
Mellum2.1 12B-A2.5B Thinking
Local OKExcellentA · 44
Best for: Agentic coding, Small machines·Pop: 76/100
Perf: est. 235 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 5080.
Gemma 4 12B (Q8)
Local OKOKB · 11
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: est. 61 tok/s, typical error 15% · first token ~1.3s
Fits in 16 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 5080.
Gemma 4 E4B (Q8)
Local OKExcellentA · 48
Best for: On-device, Mobile, Chat·Pop: 82/100
Perf: est. 141 tok/s, typical error 5% · first token ~0.8s
Fits in 16 GB VRAM with room to spare. Best for on-device, mobile, chat on RTX 5080.
NVIDIA Nemotron 3.5 Bolt E4B
Local OKExcellentA · 45
Best for: Voice assistants, Tool calling, Chat·Pop: 75/100
Perf: est. 172 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 5080.
Ministral 3 14B
Local OKExcellentA · 37
Best for: Chat, Reasoning, Vision·Pop: 74/100
Perf: est. 84 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 5080.
Qwen3.5 9B Instruct
Local OKExcellentA · 51
Best for: Quality, Coding, Reasoning·Pop: 86/100
Perf: est. 140 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 5080.
LFM2 24B-A2B Instruct
Local OKOKB · 3
Best for: Local AI agents, privacy-first tool calling, MCP workflows·Pop: 80/100
Perf: est. 255 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 5080.
Ornith 1.0 9B
Local OKExcellentA · 61
Best for: Agentic coding on small machines·Pop: 76/100
Perf: est. 140 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 5080.
Video and image models on the RTX 5080
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: the complete download is larger than 16 GB, so part of it has to sit in system RAM. Every VRAM tier: video · image.
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 · Stated10 GB VRAM · FP4 on RTX 50 (Blackwell) · Black Forest Labs blog: FLUX with NVIDIA | #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 |
| 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.
Can I run it on the RTX 5080?
Per-model verdict pages for this exact card: weights, KV cache by context length, and estimated tok/s, each derived from the same engine the recommendations above use.
- Qwen3 8B Q4_K_Mfits
- Qwen3 14B Q4_K_Mfits
- GPT-OSS 20B MXFP4fits
- Qwen3.8 27B Q4_K_Mtight
- Llama 3.3 70B Instruct Q4_K_Mno
- Qwen3.5 35B-A3B Instruct Q4_K_Mtight
- Gemma 4 26B-A4B Q4_K_Mtight
- Qwen3.5 4B Instruct Q4_K_Mfits
- Gemma 4 E4B Q4_K_Mfits
- Gemma 3 4B Instruct Q4_K_Mfits
- Llama 3.1 8B Instruct Q4_K_Mfits
- Qwen3.5 9B Instruct Q4_K_Mfits
- Gemma 4 12B Q4_K_Mfits
- Mistral Nemo 12B Q4_K_Mfits
- Qwen3.5 9B Instruct (Q8) Q8_0fits
- Gemma 4 12B (Q8) Q8_0fits
- LFM2 24B-A2B Instruct Q4_K_Mfits
- Qwen3.5 27B Instruct Q4_K_Mtight
- Qwen3.6 27B Q4_K_Mtight
- Gemma 4 31B Q4_K_Mtight
- Qwen3.6 35B-A3B Q4_K_Mno
- Qwen3.8 27B (Q8) Q8_0no
Similar GPUs for local AI
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 5080 FAQ: common questions
How much VRAM does the RTX 5080 have for LLMs?
The RTX 5080 has 16GB GDDR7 VRAM with 960 GB/s bandwidth, the highest of any 16GB consumer card. About 15.5GB is usable for models. Perfect for 14B models at Q4 with room for generous context windows.
What size LLM can I run on an RTX 5080?
Up to 14B parameter models at Q4 quantization. The 5080 runs them at est. 83 tok/s, typical error 15% (8B models hit est. 132 tok/s, typical error 5%), faster than any other 16GB card. For 32B models, you need 24GB+ VRAM: consider the RTX 4090 or 5090 instead.
Is the RTX 5080 good for local AI in 2026?
The RTX 5080 is the best 16GB card for AI speed in 2026. At est. 132 tok/s, typical error 5% on 8B models, it beats the RTX 3090 while costing less. The only downside is that 16GB limits you to 14B models; the 5090 (32GB) unlocks 70B models.
RTX 5080 vs RTX 5090 for running AI models?
The RTX 5090 (32GB) has double the VRAM and 61% more 8B speed (est. 213 tok/s, typical error 5% vs est. 132 tok/s), but costs far more at current pricing. Get the 5080 if 14B models are enough. Get the 5090 only if you need 32B-70B models.
RTX 5080 vs RTX 5070 Ti: which is better for AI?
Both have 16GB VRAM. The 5080 is 6% faster (est. 132 tok/s, typical error 5% vs est. 125 tok/s on 8B) but costs noticeably more. For most AI workloads, the 5070 Ti is the better value. The 5080 is for users who want maximum speed from 16GB.
How fast is a 27B-class model on the RTX 5080?
The RTX 5080's 16GB of VRAM cannot fit a 32B model comfortably. The largest size class it fits is 20B MoE (3.6B active), at est. 171 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
- 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.
Measure a fixed Ollama reference model on your machine. This separate, opt-in contribution does not change ModelFit's third-party SWE-Bench leaderboard; results appear on the measured-speed leaderboard after the next site update.
$ npx -y @wecko-ai/modelfit bench --submit --cleanup
Want personalized recommendations?
Use our interactive wizard to compare models across Apple Silicon and NVIDIA GPUs.