Qwen3.8 27B
Local OKOKB · 24
Best for: Coding, Agent, Vision, Long context·Pop: 95/100
Perf: est. 48 tok/s, typical error 5% · first token ~0.6s
Fits in 24 GB VRAM with room to spare. Best for coding, agent, vision, long context on RTX 3090 Ti.
Best Local AI Models for RTX 3090 Ti (24GB)
The RTX 3090 Ti has 24GB of GDDR6X VRAM at 1,008 GB/s, the same bandwidth as the RTX 4090 and more than the RTX 3090 (936 GB/s). It holds dense models up to 32B at Q4 fully in VRAM: an 8B model runs at est. 124 tok/s, typical error 5% and a 32B model at est. 37 tok/s.
Quick answer
The best local LLM for the RTX 3090 Ti is Qwen3.8 27B (est. 48 tok/s, typical error 5%) on its 24GB VRAM. It uses ~16.5GB of VRAM; the RTX 3090 Ti handles dense models up to ~32B at Q4. A 32B model at Q4 runs at est. 37 tok/s, typical error 5%. (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 24GB for weights, context, and KV-cache. The per-size table below uses that same budget. Other strong fits: Gemma 4 26B-A4B (26B, ~16GB) and Gemma 4 12B (Q8) (12B, ~12.8GB). On this card Qwen3.8 27B runs at est. 48 tok/s, typical error 5%. What it will not run: a 35B MoE (3B active) model does not fit in VRAM alone, so it spills to system RAM over PCIe and slows sharply.
$ollama run qwen3.8:27b
- Top pick
- Qwen3.8 27B
- Decode speed
- est. 48 tok/s, typical error 5%
- VRAM needed
- ~16.5 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 3090 Ti VRAM: 24GB for Local LLMs, 32B at est. 37 tok/s, https://modelfit.io/gpu/rtx-3090-ti/, updated October 2026, CC BY 4.0.
Specs
VRAM
24 GB GDDR6X
Speed (8B Q4)
est. 124 tok/stypical error 5%
Bandwidth
1008 GB/s
Architecture
Ampere
Max model size
Up to ~32B dense
Compatibility
10 excellent, 0 workable
Table 1
RTX 3090 Ti estimated tokens/sec by model size
Q4_K_M · ModelFit estimate| Model size | Est. speed | Fit on 24GB |
|---|---|---|
| 7B | est. 140 tok/stypical error 5% | Fits in VRAM |
| 14B | est. 79 tok/stypical error 15% | Fits in VRAM |
| 20B MoE (3.6B active) | est. 160 tok/stypical error 15% | Fits in VRAM |
| 32B | est. 37 tok/stypical error 5% | Fits in VRAM |
| 35B MoE (3B active) | est. 137 tok/srough estimate, offload not measured | CPU offload (slow) |
| 70B | est. 2 tok/srough estimate, offload not measured | CPU offload (slow) |
| 120B MoE (5.1B active) | est. 43 tok/srough estimate, offload not measured | CPU offload (slow) |
Fig. 1RTX 3090 Ti: estimated speed by model size
Decode speed, generic Q4_K_M model of each size at 4K context, 1008 GB/s, typical error 5-15%
Fits in 24 GBCPU offload (slow)
7B140
14B79
20B MoE (3.6B active)160
32B37
35B MoE (3B active)137
70B2
120B MoE (5.1B active)43
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 3090 Ti's 1008 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 24GB 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.
Qwen3.8 27B loads ~16.5 GB of weights; at 16k context the KV cache adds ~1.0 GB (still fits the ~22 GB usable VRAM), and at 64k it adds ~4.0 GB (still fits).
Weights 17 GBKV cache · 16k context 1.0 GBHeadroom 4.1 GBRuntime reserve 2.4 GB
The model leaves about 4.1 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 Qwen3.8 27B on your RTX 3090 Ti
One email when a new open-weight model is a better fit for 24 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 3090 Ti
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 3090 Ti VRAM for AI: what actually fits?
ModelFit budgets about 21.6GB of the 24GB for weights and context. Qwen3.6 27B at Q4 (~18GB) loads with room for context, and 32B models at Q4 fit with a shorter one. Against the RTX 3090 the gain is bandwidth only, 6% on 8B models (est. 124 tok/s, typical error 5% vs est. 117 tok/s): both cards hold the same models. A 70B model does not fit one card and spills into system RAM.
Table 2
RTX 3090 Ti vs similar GPUs
| Hardware | Memory | Speed | Bandwidth |
|---|---|---|---|
| RX 7900 XTX | 24 GB | est. 104 tok/stypical error 10% | 960 GB/s |
| RTX 3090 | 24 GB | est. 117 tok/stypical error 5% | 936 GB/s |
| RTX 3090 Ti | 24 GB | est. 124 tok/stypical error 5% | 1008 GB/s |
| RTX 4090 | 24 GB | est. 133 tok/stypical error 5% | 1008 GB/s |
| RTX 5090 | 32 GB | est. 213 tok/stypical error 5% | 1792 GB/s |
Recommended models
10 modelsGemma 4 26B-A4B
Local OKOKB · 26
Best for: Chat, Coding, Multimodal·Pop: 86/100
Perf: est. 156 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 3090 Ti.
Gemma 4 12B (Q8)
Local OKExcellentA · 41
Best for: Chat, Coding, Multimodal·Pop: 80/100
Perf: est. 58 tok/s, typical error 15% · first token ~1.8s
Fits in 24 GB VRAM with room to spare. Best for chat, coding, multimodal on RTX 3090 Ti.
Nemotron 3.5 Lightning 30B-A3B (IQ4_XS)
Local OKOKB · 13
Best for: Agentic, Coding, Long context·Pop: 78/100
Perf: est. 185 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for agentic, coding, long context on RTX 3090 Ti.
Qwen3.5 35B-A3B Instruct
Local OKOKB · 7
Best for: Reasoning, Coding, Agent scenarios·Pop: 90/100
Perf: est. 223 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for reasoning, coding, agent scenarios on RTX 3090 Ti.
LFM2 24B-A2B Instruct
Local OKExcellentA · 35
Best for: Local AI agents, privacy-first tool calling, MCP workflows·Pop: 80/100
Perf: est. 236 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for local ai agents, privacy-first tool calling, mcp workflows on RTX 3090 Ti.
Gemma 4 31B
Local OKOKB · 7
Best for: Quality, Coding, Multimodal·Pop: 84/100
Perf: est. 39 tok/s, typical error 5% · first token ~0.7s
Fits in 24 GB VRAM with room to spare. Best for quality, coding, multimodal on RTX 3090 Ti.
GLM-4.7-Flash 30B-A3B
Local OKOKB · 12
Best for: Coding, Agentic tasks, Reasoning·Pop: 78/100
Perf: est. 165 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for coding, agentic tasks, reasoning on RTX 3090 Ti.
AREX-2 27B
Local OKOKB · 20
Best for: Agents, Deep research, Coding·Pop: 60/100
Perf: est. 47 tok/s, typical error 5% · first token ~0.6s
Fits in 24 GB VRAM with room to spare. Best for agents, deep research, coding on RTX 3090 Ti.
Laguna XS 2.1
Local OKOKB · 6
Best for: Agentic coding, Long-horizon tasks·Pop: 72/100
Perf: est. 209 tok/s, typical error 15% · first token ~0.1s
Fits in 24 GB VRAM with room to spare. Best for agentic coding, long-horizon tasks on RTX 3090 Ti.
Video and image models on the RTX 3090 Ti
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. Every VRAM tier: video · image. Ranked with reasons: video models for 24 GB of VRAM · image models for 24 GB of VRAM.
Video
| Model | Fits | Quality | App |
|---|---|---|---|
| HunyuanVideo-1.5Tencent Hunyuan Community License | Runs · Stated24 GB VRAM · ComfyUI native workflow · ComfyUI docs: HunyuanVideo 1.5 | #38 of 48 in text to video4,292 votessnapshot 2026-09-22 | ComfyUI |
| LTX-2.3LTX-2 Community License Agreement | Runs · Stated24 GB VRAM · two-stage production at 720p to 1080p · 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 · Stated24 GB VRAM · LTX Desktop app, 32 GB system RAM · LTX Desktop README (Lightricks) | No independent score yet | Lightricks/LTX-2 ltx-pipelines, ComfyUI |
| MiniMax H3MiniMax H3 Community License Agreement | Runs · Stated24 GB VRAM · Diffusers int8 with streamed blocks, about 75 GB system RAM · Diffusers docs: MiniMax-H3 memory | #8 of 48 in text to video11,417 votessnapshot 2026-09-22 | stable-diffusion.cpp, ComfyUI |
| Mochi 1 previewApache License 2.0 | Runs · Stated22 GB VRAM · Diffusers bf16 variant with CPU offload · Mochi 1 model cardComplete: Official Diffusers BF16 variant40.0 GB download | #48 of 48 in text to video5,893 votessnapshot 2026-09-22 | Diffusers, 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 · Stated24 GB VRAM · official generate.py at 1280x704 with offload flags · Wan2.2-TI2V-5B model card | No independent score yet | 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 [dev]FLUX Non-Commercial License | Runs · Stated24 GB VRAM · 4-bit quantization in Diffusers · FLUX.2 dev Diffusers guide (Black Forest Labs) | #38 of 82 in text to image77,988 votessnapshot 2026-10-07 | Diffusers, ComfyUI, stable-diffusion.cpp |
| 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 4090 · 24 GB VRAM · Diffusers bf16 with model CPU offload, 1024x1024, 50 GB free system RAM required · MindrLabs Qwen-Image-2.1 README | #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 |
| 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 |
| Stable Diffusion 3.5 LargeStability AI Community License | unknownComplete: Q4_0 GGUF16.4 GB download | #81 of 82 in text to image23,396 votessnapshot 2026-10-07 | ComfyUI |
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
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 3090 Ti FAQ: common questions
How much VRAM does the RTX 3090 Ti have?
The RTX 3090 Ti has 24GB of GDDR6X VRAM at 1,008 GB/s (NVIDIA Ada whitepaper). ModelFit budgets about 21.6GB of it for weights and context, enough for dense models up to 32B at Q4 quantization.
RTX 3090 Ti vs RTX 3090 for local LLMs?
Both have 24GB, so they load the same models. The Ti streams 1,008 GB/s against 936 GB/s, which makes it 6% faster on 8B models (est. 124 tok/s, typical error 5% vs est. 117 tok/s). Pick on price: the extra speed is small.
What size LLM can I run on an RTX 3090 Ti?
Dense models up to 32B at Q4 fully in VRAM, such as Qwen3.6 27B (~18GB) or a 32B model at est. 37 tok/s, typical error 5%. A 70B model at Q4 (~42GB) is too large for one card.
RTX 3090 Ti vs RTX 4090 for local AI?
Both have 24GB at 1,008 GB/s, so they hold the same models and decode at similar speeds (est. 124 tok/s, typical error 5% vs est. 133 tok/s on 8B). The 4090 has about twice the tensor throughput, so it reads long prompts faster.
How fast is a 27B-class model like Qwen3.8 27B on the RTX 3090 Ti?
ModelFit estimates a 32B model on the RTX 3090 Ti at Q4_K_M runs at est. 37 tok/s, typical error 5%. The current 27B-class pick in the catalog is Qwen3.8 27B (ollama run qwen3.8:27b).
Related guides & benchmarks
More local AI guides
- Best LLM for 24GB VRAMEvery 24 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.