Skip to content
Local AI Wizard
Your machine
Theme

RTX 5070 Ti · 16 GB VRAM

NVIDIA · RTX 5070 Ti

Graphics cards16 GB VRAM

125 text models run well · 896 GB/s bandwidth

Models on this hardware

Start here

Text

Within 16 GB: 125 run well · 21 reviewed · 36 more run slowly (11 of them run well with 24 GB)

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.

Rows run well first, then tight fits; within each, reviewed models, then Hub downloads over 30 days, not quality. Sort a score column to rank by that source alone. All 11 text models that would run well with 24 GB but not here already run here, slowly.
Text models that run well within 16 GB
Granite 4.2 3BIBM · 3B · LMArena 1,290Q4 · published params · source1,290 ± 12~5 GBQ4 · published paramssource
Granite 4.2 8BIBM · 8B · LMArena 1,290Q4 · published params · source1,290 ± 11~8 GBQ4 · published paramssource
Phi-4Microsoft · 14B · LMArena 1,256Q4 · published params · source1,256 ± 5~13 GBQ4 · published paramssource
Llama 3.1 8B InstructMeta · 8B · LMArena 1,211Q4 · published params · source1,211 ± 4~8 GBQ4 · published paramssource
Llama 3.2 3B InstructMeta · 3.21B · LMArena 1,167Q4 · published params · source1,167 ± 8~5 GBQ4 · published paramssource
Llama 3.2 1B InstructMeta · 1.23B · LMArena 1,111Q4 · published params · source1,111 ± 8~3 GBQ4 · published paramssource
Qwen3 14BAlibaba · 14.8BQ4 · published params · sourceNot on the LMArena text leaderboard of 2026-10-22~14 GBQ4 · published paramssource
Qwen3 8BAlibaba · 8.2BQ4 · published params · sourceNot on the LMArena text leaderboard of 2026-10-22~9 GBQ4 · published paramssource
8 of 125

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.

Reviewed text models that run on RTX 5070 Ti · 16 GB VRAM
Qwen3 14BAlibaba · 14.8BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~14 GB9.0Apache 2.010.4M
Qwen3 8BAlibaba · 8.2BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~9 GB5.0Apache 2.08.1M
Qwen3.5 9BAlibaba · 9BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~9 GB5.7Apache 2.06.1M
Llama 3.1 8B InstructMeta · 8BNot in LiveBench 2026-06-251,211~8 GB4.9Llama 3.14.8M
Qwen3 1.7BAlibaba · 1.7BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~4 GB1.1Apache 2.04.1M
Gemma 4 E4BGoogle · 8BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~8 GB5.0Apache 2.04M
Llama 3.2 1B InstructMeta · 1.2BNot in LiveBench 2026-06-251,111~3 GB0.8Llama 3.23.5M
Gemma 4 E2BGoogle · 5.1BNot in LiveBench 2026-06-25Not on the LMArena text leaderboard of 2026-10-22~6 GB3.1Apache 2.03M
8 of 37
Also runs slowly20 models

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.

6 of 20

Browse Text models

Adjust the estimate: quantization, context, reserveloads the live catalog

Vision

Within 16 GB: 125 run well · 4 reviewed · 17 more run slowly (8 of them run well with 24 GB)

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.

Rows run well first, then tight fits; within each, reviewed models, then Hub downloads over 30 days, not quality. Sort a score column to rank by that source alone. All 8 vision models that would run well with 24 GB but not here already run here, slowly.
Vision models that run well within 16 GB
Molmo2-8Ballenai · 8.7B · Image to text · LMArena 1,112Q4 · Hugging Face count · source1,112 ± 20~9 GBQ4 · Hugging Face countsource
Qwen2-VL-7B-InstructQwen · 8.29B · Image to text · LMArena 1,031Q4 · Hugging Face count · source1,031 ± 11~9 GBQ4 · Hugging Face countsource
Molmo-7B-D-0924allenai · 8B · Image to text · LMArena 994Q4 · Hugging Face count · source994 ± 14~8 GBQ4 · Hugging Face countsource
Llama-3.2-11B-Vision-Instructmeta-llama · 10.7B · Image to text · LMArena 993Q4 · Hugging Face count · source993 ± 11~11 GBQ4 · Hugging Face countsource
MiniCPM-V-2_6openbmb · 8.1B · Image to text · LMArena 962Q4 · Hugging Face count · source962 ± 16~9 GBQ4 · Hugging Face countsource
Phi-3.5-vision-instructmicrosoft · 4.1B · Image to text · LMArena 919Q4 · Hugging Face count · source919 ± 16~6 GBQ4 · Hugging Face countsource
Gemma 4 12BGoogle · 11.95B · Image to textQ4 · published params · sourceNot on the LMArena vision leaderboard of 2026-10-19~11 GBQ4 · published paramssource
Qwen3.5 9BAlibaba · 9B · Image to textQ4 · published params · sourceNot on the LMArena vision leaderboard of 2026-10-19~9 GBQ4 · published paramssource
8 of 125

LMArena ratings in this table: LMArena 2026-10-19 · collected 2026-10-25 · dataset · CC BY 4.0

Also runs slowly17 models

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.

6 of 17

12 more vision models in the catalog have no usable memory figure or parameter count for this profile. Browse Vision models

Image

30 within 16 GB · 23 more with 24 GB

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.

Image models within 16 GB
Stable Diffusion XL 1.0Stability AI · 2.6B · Text to imagePlatform path documented · stable-diffusion.cpp · source0.55Not on the LMArena text-to-image leaderboard of 2026-10-07Documented path without a memory figurePlatform path documentedstable-diffusion.cpp · source
Stable Diffusion 1.5Stability AI · 0.98B · Text to imagePlatform path documented · stable-diffusion.cpp · source0.43Not on the LMArena text-to-image leaderboard of 2026-10-07Documented path without a memory figurePlatform path documentedstable-diffusion.cpp · source
TRELLIS-image-largemicrosoft · 1.2B · 3DPublisher minimum · CUDA · sourceNo comparable published resultNot reviewed for LMArena~16 GBPublisher minimumCUDA · source
FLUX.2 [klein] 4BBlack Forest Labs · 4B · Text to image · LMArena 1,030Publisher minimum · CUDA · sourceNo comparable published result1,030 ± 3~13 GBPublisher minimumCUDA · source
FLUX.2 [klein] Base 4BBlack Forest Labs · 4B · Text to imagePublisher minimum · CUDA · sourceNo comparable published resultArena rated flux-2-klein-4b; this entry is the undistilled Base checkpoint~13 GBPublisher minimumCUDA · source
BagelByteDance · 14B · Text to image · LMArena 898Conditional path · CUDA · sourceNo comparable published result898 ± 6~12 GBConditional pathCUDA · source
Stable Diffusion 3.5 LargeStability AI · 8B · Text to image · LMArena 938Conditional path · CUDA · sourceNo comparable published result938 ± 5~11 GBConditional pathCUDA · source
Stable Diffusion 3.5 MediumStability AI · 2.47B · Text to imageConditional path · sourceNo comparable published resultNot on the LMArena text-to-image leaderboard of 2026-10-07~10 GBConditional pathsource
8 of 30

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.

6 of 23

17 more image models in the catalog have no usable memory figure or parameter count for this profile. Browse Image models

Video

27 within 16 GB · 3 more with 24 GB
Video models within 16 GB
CogVideoX-5b-I2Vzai-org · 5.6B · Image to video16-bit est. · params · source~16 GB16-bit est. · paramssource
CogVideoX1.5-5Bzai-org · 5.6B · Text to video16-bit est. · params · source~16 GB16-bit est. · paramssource
CogVideoX1.5-5B-I2Vzai-org · 5.6B · Image to video16-bit est. · params · source~16 GB16-bit est. · paramssource
SANA-Video_2.0_5B_720pEfficient-Large-Model · 4.5B · Text to video16-bit est. · params · source~13 GB16-bit est. · paramssource
Kandinsky-6.0-Lite-distill-5s-Diffuserskandinskylab · 3.2B · Image to video16-bit est. · params · source~10 GB16-bit est. · paramssource
Allegrorhymes-ai · 2.8B · Text to video16-bit est. · params · source~9 GB16-bit est. · paramssource
Wan2.1 T2V 1.3BWan-AI · 1.3B · Text to videoPlatform path documented · Wan2.1 / PyTorch · source~8 GBPlatform path documentedWan2.1 / PyTorch · source
Wan2.1-T2V-1.3B-DiffusersWan-AI · 1.4B · Text to videoConditional path · CUDA · source~8 GBConditional pathCUDA · source
8 of 27

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.

20 more video models in the catalog have no usable memory figure or parameter count for this profile. Browse Video models

Audio

116 within 16 GB · 12 more with 24 GB

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.

Audio models within 16 GB
Whisper large-v3-turboOpenAI · 0.81B · Speech to textPlatform path documented · whisper.cpp / CUDA · source7.83Documented path without a memory figurePlatform path documentedwhisper.cpp / CUDA · source
Whisper large-v3OpenAI · 1.55B · Speech to textPlatform path documented · whisper.cpp / CUDA · source7.44Documented path without a memory figurePlatform path documentedwhisper.cpp / CUDA · source
GaussianMindBAAI · 5.6B · Speech to text16-bit est. · params · sourceNo comparable published result~16 GB16-bit est. · paramssource
Higgs Audio V3 TTSBoson AI · 4B · Text to speechPublisher minimum · CUDA · sourceNo comparable published result~16 GBPublisher minimumCUDA · source
higgs-tts-2-3b-basebosonai · 5.8B · Text to speech16-bit est. · params · sourceNo comparable published result~16 GB16-bit est. · paramssource
Maya1Maya Research · 3B · Text to speechPublisher minimum · CUDA · sourceNo comparable published result~16 GBPublisher minimumCUDA · source
MusicGen SmallMeta FAIR · 0.3B · MusicPlatform path documented · AudioCraft / PyTorch · sourceNo comparable published result~16 GBPlatform path documentedAudioCraft / PyTorch · source
Phi-4-multimodal-instructmicrosoft · 5.6B · Speech to text16-bit est. · params · sourceNo comparable published result~16 GB16-bit est. · paramssource
8 of 116

With 24 GB you also run… 12

The next memory size on this ladder. Figures are the requirement on this hardware type.

6 of 12

33 more audio models in the catalog have no usable memory figure or parameter count for this profile. Browse Audio models

Embeddings

89 run well within 16 GB · 1 more with 24 GB

Embeddings · MTEB (Multilingual, v2) · mean(task)

0–100 · higher is better

  • Color = publisher

Source: MTEB results (curated rows, sourced per model)

Embeddings models that run well within 16 GB
Qwen3-Embedding 8BAlibaba (Qwen)8-bit · published params · source70.6~12 GB8-bit · published paramssource
Qwen3-Embedding 4BAlibaba (Qwen)8-bit · published params · source69.5~7 GB8-bit · published paramssource
Harrier OSS v1 (0.6B)Microsoft8-bit · published params · source69.0~3 GB8-bit · published paramssource
Qwen3-Embedding 0.6BAlibaba (Qwen)8-bit · published params · source64.3~3 GB8-bit · published paramssource
Multilingual E5 Large Instructintfloat8-bit · published params · source63.2~3 GB8-bit · published paramssource
EmbeddingGemma 2Google8-bit · published params · source61.4~3 GB8-bit · published paramssource
EmbeddingGemma (300M)Google8-bit · published params · source61.1~3 GB8-bit · published paramssource
BGE-M3BAAI8-bit · published params · source59.6~3 GB8-bit · published paramssource
8 of 89

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.

Browse Embeddings models

Runs well within 16 GB of VRAM · planning estimates, not measured speed bundled catalog copy · 1,777 repositories · checked

Measurements

Published tokens per second, exactly as measured; never estimated.

Published measurements on this hardware
gpt-oss 20B Q4_K_Munsloth/gpt-oss-20b-GGUF/gpt-oss-20b-Q4_K_M…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)236.89 ± 2.18Q4_K_M
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
gpt-oss 20B Q8_0unsloth/gpt-oss-20b-GGUF/gpt-oss-20b-Q8_0.g…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)223.82 ± 2.47Q8_0
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
gpt-oss-20b MXFP4gpt-oss-20b-mxfp4.gguf (ggml-org/gpt-oss-20…NVIDIA GeForce RTX 5070 Ti · 16 GB189.45 ± 0.09MXFP4 (native gpt-oss weights)
0 / 128 · Single llama-bench process;…
llama.cpp CUDA
llama.cpp commit a094f38 (6210)
Flash Attention enabled
llama.cpp gpt-oss guide · maintainer result
2026-10-20
Llama 2 7B Q4_0Llama 2 7B Q4_0 (scoreboard model; the scor…NVIDIA GeForce RTX 5070 Ti · 16 GB GDDR7182.43 ± 0.09Q4_0
0 / 128 · Single llama-bench process;…
CUDA
llama.cpp commit 933414c
Flash Attention enabled
llama.cpp CUDA scoreboard · community result
2026-10-25
Qwen3.5 9B Q4_K_Munsloth/Qwen3.5-9B-GGUF/Qwen3.5-9B-Q4_K_M.g…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)126.03 ± 0.80Q4_K_M
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
Gemma 4 12B Q4_K_Munsloth/gemma-4-12b-it-GGUF/gemma-4-12b-it-…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)89.15 ± 0.31Q4_K_M
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
Qwen3.5 9B Q8_0unsloth/Qwen3.5-9B-GGUF/Qwen3.5-9B-Q8_0.ggufNVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)86.39 ± 0.24Q8_0
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
Gemma 4 12B Q8_0unsloth/gemma-4-12b-it-GGUF/gemma-4-12b-it-…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)58.41 ± 0.18Q8_0
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25
Mistral Small 3.2 24B Q4_K_Munsloth/Mistral-Small-3.2-24B-Instruct-2506…NVIDIA GeForce RTX 5070 Ti · 16 GB (the card's only memory size; rented host)52.85 ± 0.07Q4_K_M
0 / 128 · Single llama-bench process
llama.cpp CUDA
llama.cpp commit 3e4dfa8f3 (b11658)
Flash Attention enabled
Data: Local LLM Advisor (localllmadvisor.com) · CC BY 4.0
2026-10-25

Specifications

11 sourced fields; the source is named once per group, and a field from another source links to it. Official specifications

GPU VRAM16 GBDedicated · system RAM does not add · source
Memory bandwidth896 GB/sManufacturer spec · not model speed · source
Launch price (MSRP)USD 749.00US · launched · source · not a tracked offer

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.

Selected configuration

Latest check · Manufacturer source

Memory capacity
16 GB
Memory type
16 GB GDDR7 · 256-bit
Chip / processor
Blackwell · RTX 5070 Ti
Memory bandwidth
896 GB/s
CUDA cores
8,960
GPU power (reference specification)
300 W
Memory interface
256 bit
CUDA compute capability
12.0

Family specification

Latest check · Manufacturer source

Display connectors (reference design; OEM may vary)
3 × DisplayPort 2.1b · HDMI 2.1b
Release date
2025-02-20
Introduced
2025

Desktop RTX 5070 Ti with 16 GB GDDR7. Capacity alone does not establish local throughput; verify runtime and checkpoint format.

Photo: ASUS partner-board photo of a GeForce RTX 5070 Ti · other partner boards vary. · Manufacturer source

Memory bandwidth is not measured model speed; verify runtime and checkpoint compatibility. Hardware specification reviewed: 2026-10-09.

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.