Your rig

GeForce RTX 5070 Ti · 16 GB

16 GB system RAM · Blackwell · 896 GB/s · 14.4 GB usable VRAM
Edit hardware Export as JSON

Context length

KV cache at f16, 32K tokens

Quality floor
Q8_0 — near-lossless Q6_K Q5_K_M Q4_K_M — recommended Q3_K_M Q2_K — damaged
KV cache type
System RAM (GB)
Runtime
Ollamallama.cppLM StudiovLLM

4 of 5 runtimes drive this device.

Allow CPU offload Include partial-offload fits

Offloaded layers run at 60 GB/s host bandwidth — expect single-digit tokens per second.

Which one should I actually run?

Best models for your GeForce RTX 5070 Ti

Updated 8 Oct 2026 1 new this month

What do you want to do? Each row is the best-ranked model on that use case's shortlist that fits at Q3_K_M and 32K. The ordering is editorial; the sizes and speeds are computed. How the shortlists work.

RecommendationModelWhyQuant · totalTok/s est.
Best overall Gemma 4 12B
Google · Apache 2.0
The 12 GB generalist: text, image and audio in, 140+ languages. Q3_K_M · 8.4 GB 99 Runs great
Best for coding Qwen3.5 9B
Alibaba · Apache 2.0
The best coding model for 8–12 GB cards. Q3_K_M · 6.0 GB 123 Runs great
Best reasoning gpt-oss 20B
OpenAI · Apache 2.0
Low/medium/high reasoning effort in 13 GB. MXFP4 · 12.2 GB 112 Runs great
Best for writing Gemma 4 12B
Google · Apache 2.0
The 12 GB writing pick; 140+ languages. Q3_K_M · 8.4 GB 99 Runs great
Best vision Gemma 4 26B-A4B
Google · Apache 2.0
Vision at MoE speed in 18 GB. Q3_K_M · 14.1 GB 121 Tight fit
Best for agents gpt-oss 20B
OpenAI · Apache 2.0
Harmony-format tool calling, native 4-bit. MXFP4 · 12.2 GB 112 Runs great
Best translation Gemma 4 12B
Google · Apache 2.0
Broad language coverage on a 12 GB card. Q3_K_M · 8.4 GB 99 Runs great
Fastest good model Gemma 4 26B-A4B
Google · Apache 2.0
3.8B active and a small file — the quickest Gemma 4 with vision. Q3_K_M · 14.1 GB 121 Tight fit
Best long context Qwen3.5 9B
Alibaba · Apache 2.0
262K native in 8–12 GB. Q3_K_M · 9.0 GB
at 128K context
123 Runs great

Everything that fits

Runs great

35 of 80 · fits with headroom for a longer prompt
ModelParamsQuantWeights +KV 32KTotal VRAM headroom Tok/sRuntime
gpt-oss 20B 20.9B MoE / 3.6B act MXFP4 10.8 GB 0.75 GB 12.2 GB
2.2 GB free
112 Ollama Run
Phi-4 14B 14.7B Q3_K_M 6.7 GB 3.13 GB 10.4 GB
4.0 GB free
81 Ollama Run
Ministral 3 14B 13.9B Q3_K_M 6.3 GB 5.00 GB 11.9 GB
2.5 GB free
86 Ollama Run
Gemma 3 12B 12.2B Q3_K_M 5.6 GB 2.31 GB 8.5 GB
5.9 GB free
98 Ollama Run
Mistral NeMo 12B 12.2B Q3_K_M 5.6 GB 5.00 GB 11.2 GB
3.2 GB free
98 Ollama Run
Gemma 4 12B 12B Q3_K_M 5.5 GB 2.31 GB 8.4 GB
6.0 GB free
99 Ollama Run
Qwen3.5 9B 9.65B Q3_K_M 4.4 GB 1.00 GB 6.0 GB
8.4 GB free
123 Ollama Run
Ornith 1.5 9B 9.41B Q3_K_M 4.3 GB 1.00 GB 5.9 GB
8.5 GB free
127 Ollama Run
Ministral 3 8B 8.92B Q3_K_M 4.1 GB 4.25 GB 8.9 GB
5.5 GB free
134 Ollama Run
Granite 4.1 8B 8.79B Q3_K_M 4.0 GB 5.00 GB 9.6 GB
4.8 GB free
136 Ollama Run
LFM2.5 8B-A1B 8.47B MoE / 1.5B act Q3_K_M 3.9 GB 0.38 GB 4.8 GB
9.6 GB free
306 Ollama Run
Fara 7B 8.29B Q3_K_M 3.8 GB 1.75 GB 6.1 GB
8.3 GB free
144 llama.cpp Run
Qwen3 8B 8.19B Q3_K_M 3.7 GB 4.50 GB 8.8 GB
5.6 GB free
145 Ollama Run
Llama 3.1 8B Instruct 8.03B Q3_K_M 3.7 GB 4.00 GB 8.3 GB
6.1 GB free
148 Ollama Run
Gemma 4 E4B 8.0B Q3_K_M 3.6 GB 0.47 GB 4.7 GB
9.7 GB free
149 Ollama Run
Ling 3.0 Tiny 7.9B-A1.3B 7.9B MoE / 1.3B act Q3_K_M 3.6 GB 0.21 GB 4.4 GB
10.0 GB free
353 llama.cpp Run
DeepSeek-R1-Distill-Qwen 7B 7.62B Q3_K_M 3.5 GB 1.75 GB 5.8 GB
8.6 GB free
156 Ollama Run
Qwen2.5-Coder 7B 7.62B Q3_K_M 3.5 GB 1.75 GB 5.8 GB
8.6 GB free
156 Ollama Run
Olmo 3 7B Instruct 7.3B Q3_K_M 3.3 GB 5.50 GB 9.4 GB
5.0 GB free
163 Ollama Run
Mistral 7B Instruct v0.3 7.25B Q3_K_M 3.3 GB 4.00 GB 7.9 GB
6.5 GB free
164 Ollama Run
Gemma 4 E2B 5.1B Q3_K_M 2.3 GB 0.23 GB 3.2 GB
11.2 GB free
234 Ollama Run
Qwen3.5 4B 4.66B Q3_K_M 2.1 GB 1.00 GB 3.7 GB
10.7 GB free
256 Ollama Run
Gemma 3 4B 4.3B Q3_K_M 2.0 GB 0.86 GB 3.4 GB
11.0 GB free
277 Ollama Run
Qwen3 4B 4.02B Q3_K_M 1.8 GB 4.50 GB 6.9 GB
7.5 GB free
296 Ollama Run
Ministral 3 3B 3.85B Q3_K_M 1.8 GB 3.25 GB 5.6 GB
8.8 GB free
310 Ollama Run
Phi-4-mini 3.8B 3.84B Q3_K_M 1.7 GB 4.00 GB 6.3 GB
8.1 GB free
310 Ollama Run
Granite 4.1 3B 3.4B Q3_K_M 1.5 GB 2.50 GB 4.6 GB
9.8 GB free
350 Ollama Run
Llama 3.2 3B Instruct 3.21B Q3_K_M 1.5 GB 3.50 GB 5.6 GB
8.8 GB free
371 Ollama Run
LFM2.5 2.6B 2.7B Q3_K_M 1.2 GB 0.50 GB 2.3 GB
12.1 GB free
441 llama.cpp Run
Qwen3.5 2B 2.27B Q3_K_M 1.0 GB 0.38 GB 2.0 GB
12.4 GB free
525 Ollama Run
Qwen3 1.7B 1.72B Q3_K_M 0.8 GB 3.50 GB 4.9 GB
9.5 GB free
693 Ollama Run
Llama 3.2 1B Instruct 1.24B Q3_K_M 0.6 GB 1.00 GB 2.2 GB
12.2 GB free
961 Ollama Run
Gemma 3 1B 1.0B Q3_K_M 0.5 GB 0.14 GB 1.2 GB
13.2 GB free
1192 Ollama Run
Qwen3.5 0.8B 0.87B Q3_K_M 0.4 GB 0.38 GB 1.4 GB
13.0 GB free
1370 Ollama Run
Qwen3 0.6B 0.6B Q3_K_M 0.3 GB 3.50 GB 4.4 GB
10.0 GB free
1986 Ollama Run

Runs — tight

4 models · fits at 32K, drop context or quant to go longer
ModelParamsQuantWeights +KV 32KTotal VRAM headroom Tok/sRuntime
Gemma 4 26B-A4B 26.5B MoE / 3.8B act Q3_K_M 12.1 GB 1.45 GB 14.1 GB
0.3 GB free
121 llama.cpp Run
Qwen3 14B 14.8B Q3_K_M 6.7 GB 5.00 GB 12.3 GB
2.1 GB free
81 llama.cpp Run
DeepSeek-R1-Distill-Qwen 14B 14.8B Q3_K_M 6.7 GB 6.00 GB 13.3 GB
1.1 GB free
81 llama.cpp Run
Qwen2.5-Coder 14B 14.8B Q3_K_M 6.7 GB 6.00 GB 13.3 GB
1.1 GB free
81 llama.cpp Run

Won't fit in VRAM

41 models · some are possible with CPU offload
ModelParamsQuantWeights +KV 32KTotal Over budget Tok/sRuntime
Qwen3.6 35B-A3B 35.9B MoE / 3.3B act Q3_K_M 16.3 GB 0.63 GB 17.6 GB +3.2 GB ~38 llama.cpp Why
Ornith 1.5 35B-A3B 35.9B MoE / 3.3B act Q3_K_M 16.3 GB 0.63 GB 17.6 GB +3.2 GB ~38 llama.cpp Why
LLM-jp 4 33B Thinking 33.2B Q3_K_M 15.1 GB 8.00 GB 23.7 GB +9.3 GB ~3.7 llama.cpp Why
Qwen3 32B 32.8B Q3_K_M 14.9 GB 8.00 GB 23.5 GB +9.1 GB ~3.8 llama.cpp Why
DeepSeek-R1-Distill-Qwen 32B 32.8B Q3_K_M 14.9 GB 8.00 GB 23.5 GB +9.1 GB ~3.8 llama.cpp Why
Qwen2.5-Coder 32B 32.8B Q3_K_M 14.9 GB 8.00 GB 23.5 GB +9.1 GB ~3.8 llama.cpp Why
Olmo 3.1 32B Instruct 32.2B Q3_K_M 14.7 GB 2.75 GB 18.0 GB +3.6 GB ~8.4 llama.cpp Why
Nemotron 3.5 Lightning 30B-A3B 31.6B MoE / 3.2B act Q3_K_M 14.4 GB 0.19 GB 15.2 GB +0.8 GB ~82 llama.cpp Why
Gemma 4 31B 31.3B Q3_K_M 14.2 GB 5.78 GB 20.6 GB +6.2 GB ~5.4 llama.cpp Why
GLM-4.7-Flash 30B-A3B 31.2B MoE / 3B act Q3_K_M 14.2 GB 1.65 GB 16.5 GB +2.1 GB ~51 llama.cpp Why
Qwen3 30B-A3B 30.5B MoE / 3.3B act Q3_K_M 13.9 GB 3.00 GB 17.5 GB +3.1 GB ~34 llama.cpp Why
Qwen3 Coder 30B-A3B 30.5B MoE / 3.3B act Q3_K_M 13.9 GB 3.00 GB 17.5 GB +3.1 GB ~34 llama.cpp Why
Muse Glimmer 30B 29.8B Q3_K_M 13.6 GB 0.48 GB 14.6 GB +0.2 GB ~32 llama.cpp Why
Granite 4.1 30B 28.9B Q3_K_M 13.2 GB 8.00 GB 21.8 GB +7.4 GB ~4.7 llama.cpp Why
Qwen3.8 27B 27.8B Q3_K_M 12.7 GB 2.00 GB 15.3 GB +0.9 GB ~22 llama.cpp Why
Qwen3.6 27B 27.8B Q3_K_M 12.7 GB 2.00 GB 15.3 GB +0.9 GB ~22 llama.cpp Why
Gemma 3 27B 27.4B Q3_K_M 12.5 GB 2.91 GB 16.0 GB +1.6 GB ~16 llama.cpp Why
Devstral Small 2 24B 24B Q3_K_M 10.9 GB 5.00 GB 16.5 GB +2.1 GB ~13 llama.cpp Why
Mistral Small 3.2 24B 23.6B Q3_K_M 10.7 GB 5.00 GB 16.3 GB +1.9 GB ~14 llama.cpp Why
Kimi K3 2.8T-A104B 2780B MoE / 104B act Q3_K_M 1265.4 GB 0.84 GB 1266.9 GB +1252.5 GB ~0.3 llama.cpp Why
Qwen3.8 2.4T-A95B 2446B MoE / 95B act Q3_K_M 1113.4 GB 2.88 GB 1116.9 GB +1102.5 GB ~0.3 llama.cpp Why
DeepSeek V4 Pro 1.6T-A49B 1650B MoE / 49B act Q3_K_M 751.1 GB 2.14 GB 753.8 GB +739.4 GB ~0.6 llama.cpp Why
Kimi K2.6 1T-A32B 1027B MoE / 32B act Q3_K_M 467.5 GB 2.14 GB 470.2 GB +455.8 GB ~1.0 llama.cpp Why
GLM-5.2 744B-A40B 753B MoE / 40B act Q3_K_M 342.8 GB 2.74 GB 346.1 GB +331.7 GB ~0.8 llama.cpp Why
GLM-5.3 744B-A40B 753B MoE / 40B act Q3_K_M 342.8 GB 2.74 GB 346.1 GB +331.7 GB ~0.8 llama.cpp Why
DeepSeek-R1 671B 671B MoE / 37B act Q3_K_M 305.4 GB 2.14 GB 308.2 GB +293.8 GB ~0.9 llama.cpp Why
DeepSeek V4.1 Flash 552B-A16B New 552B MoE / 16B act Q3_K_M 251.3 GB 0.15 GB 252.0 GB +237.6 GB ~2.0 vLLM Why
Qwen3.5 397B-A17B 403B MoE / 17B act Q3_K_M 183.4 GB 0.94 GB 185.0 GB +170.6 GB ~1.9 llama.cpp Why
Ornith 1.5 397B-A17B 397B MoE / 17B act Q3_K_M 180.7 GB 0.94 GB 182.2 GB +167.8 GB ~1.9 llama.cpp Why
GLM-5.3-Flash 320B-A18B 321B MoE / 18B act Q3_K_M 146.1 GB 0.34 GB 147.1 GB +132.7 GB ~1.9 llama.cpp Why
DeepSeek V4 Flash 284B-A13B 284B MoE / 13B act Q3_K_M 129.3 GB 1.51 GB 131.4 GB +117.0 GB ~2.6 llama.cpp Why
Qwen3 235B-A22B 235B MoE / 22B act Q3_K_M 107.0 GB 5.88 GB 113.4 GB +99.0 GB ~1.5 llama.cpp Why
Qwen3.8-Flash-Next 180B-A6B 180B MoE / 6B act Q3_K_M 81.9 GB 0.75 GB 83.3 GB +68.9 GB ~6.0 llama.cpp Why
Qwen3.5 122B-A10B 125B MoE / 10B act Q3_K_M 56.9 GB 0.75 GB 58.2 GB +43.8 GB ~3.9 llama.cpp Why
Nemotron 3 Super 120B-A12B 124B MoE / 12B act Q3_K_M 56.4 GB 0.25 GB 57.3 GB +42.9 GB ~3.3 llama.cpp Why
Ling 3.0 Flash 124B-A5B 124B MoE / 5.1B act Q3_K_M 56.4 GB 0.25 GB 57.3 GB +42.9 GB ~7.8 llama.cpp Why
Mistral Small 4 119B-A6B 119B MoE / 6.5B act Q3_K_M 54.2 GB 0.70 GB 55.5 GB +41.1 GB ~6.1 llama.cpp Why
gpt-oss 120B 117B MoE / 5.1B act MXFP4 60.5 GB 1.13 GB 62.2 GB +47.8 GB ~6.6 llama.cpp Why
Llama 4 Scout 109B-A17B 109B MoE / 17B act Q3_K_M 49.6 GB 2.63 GB 52.8 GB +38.4 GB ~2.3 llama.cpp Why
Llama 3.3 70B Instruct 70.6B Q3_K_M 32.1 GB 10.00 GB 42.7 GB +28.3 GB ~1.3 llama.cpp Why
DeepSeek-R1-Distill-Llama 70B 70.6B Q3_K_M 32.1 GB 10.00 GB 42.7 GB +28.3 GB ~1.3 llama.cpp Why

Totals = quantised weights + f16 KV cache at 32K tokens + 0.6 GB runtime overhead, against 14.4 GB usable VRAM (16 GB less display and driver reserve). Tokens per second are estimated from 896 GB/s peak memory bandwidth at batch 1 — not measured. Read the method.


Same memory, different speed

10 of the 19 devices have the same 14.4 GB usable, so they run exactly the models above. Generation speed scales with memory bandwidth.

DeviceBandwidthSpeed vs yours
GeForce RTX 5080
Blackwell · 2025
960 GB/s 1.07×
CMP 100-210
Volta · 2021
829 GB/s 0.93×
GeForce RTX 4080 Super
Ada Lovelace · 2024
736 GB/s 0.82×
Tesla P100 16 GB
Pascal · 2016
732 GB/s 0.82×
GeForce RTX 4080
Ada Lovelace · 2022
717 GB/s 0.80×
GeForce RTX 4070 Ti Super
Ada Lovelace · 2024
672 GB/s 0.75×
Radeon RX 9070 XT
RDNA 4 · 2025
645 GB/s 0.72×
Radeon RX 9070
RDNA 4 · 2025
645 GB/s 0.72×
Radeon RX 7800 XT
RDNA 3 · 2023
624 GB/s 0.70×
GeForce RTX 4090 Laptop
Ada Lovelace · 2023
576 GB/s 0.64×

Best LLM for 16 GB VRAM, with every card in the class →

What more memory unlocks

At Q4_K_M and 8K context this device fits 38 of 80 models. Each step up, and the largest models it adds:

Every device ranked by what it can run →