NVIDIA · Ampere · 12 GB VRAM tier
Best AI Models for NVIDIA GeForce RTX 3080 Ti (12 GB)
Runs 7–9B models well; 14B only just fits.
| # | Model | Speed | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 182 t/s | 182 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Gemma 3 12B ITGoogle Q4_K_M · 10.9 GB · 73.8 t/s | 73.8 tok/s | Q4_K_M · 10.9 GB | |
| 4 | Gemma 2 9B ITGoogle Q4_K_M · 7.3 GB · 97.2 t/s | 97.2 tok/s | Q4_K_M · 7.3 GB | |
| 5 | Q4_K_M · 6.5 GB · 95.5 t/s | 95.5 tok/s | Q4_K_M · 6.5 GB |
| # | Model | Speed | Coding Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Q4_K_M · 7.9 GB · 112 t/s | 112 tok/s | Q4_K_M · 7.9 GB | |
| 2 | Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | overall | Q4_K_M · 6.6 GB |
| 3 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 182 t/s | 182 tok/s | overall | Q4_K_M · 3.4 GB |
| 4 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 72.1 t/s | 72.1 tok/s | Q4_K_M · 9.0 GB | |
| 5 | Q4_K_M · 6.5 GB · 95.5 t/s | 95.5 tok/s | Q4_K_M · 6.5 GB |
| # | Model | Speed | Reasoning Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3 8BAlibaba Q4_K_M · 7.6 GB · 107 t/s | 107 tok/s | Q4_K_M · 7.6 GB | |
| 3 | Q4_K_M · 7.9 GB · 112 t/s | 112 tok/s | Q4_K_M · 7.9 GB | |
| 4 | Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 209 t/s | 209 tok/s | Q4_K_M · 5.0 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 72.1 t/s | 72.1 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed | Long context Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 182 t/s | 182 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Q4_K_M · 7.9 GB · 112 t/s | 112 tok/s | Q4_K_M · 7.9 GB | |
| 4 | Mistral Nemo Instruct 2407Mistral AI Q4_K_M · 11.0 GB · 73.5 t/s | 73.5 tok/s | Q4_K_M · 11.0 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 72.1 t/s | 72.1 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed ↓ | llmrun Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 4BAlibabaHow to run → Q4_K_M · 3.4 GB · 182 t/s | 182 tok/s | Q4_K_M · 3.4 GB | |
| 2 | Qwen3 8BAlibaba Q4_K_M · 7.6 GB · 107 t/s | 107 tok/s | Q4_K_M · 7.6 GB | |
| 3 | Q4_K_M · 6.5 GB · 95.5 t/s | 95.5 tok/s | Q4_K_M · 6.5 GB | |
| 4 | Qwen3.5 9BAlibaba Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | Q4_K_M · 6.6 GB | |
| 5 | Gemma 4 12B ITGoogle Q4_K_M · 9.0 GB · 72.1 t/s | 72.1 tok/s | Q4_K_M · 9.0 GB |
| # | Model | Speed | Vision Score | Quant · VRAM |
|---|---|---|---|---|
| 1 | Qwen3.5 9BAlibabaHow to run → Q4_K_M · 6.6 GB · 93.2 t/s | 93.2 tok/s | Q4_K_M · 6.6 GB | |
| 2 | Qwen3.5 4BAlibaba Q4_K_M · 3.4 GB · 182 t/s | 182 tok/s | Q4_K_M · 3.4 GB | |
| 3 | Gemma 3 12B ITGoogle Q4_K_M · 10.9 GB · 73.8 t/s | 73.8 tok/s | Q4_K_M · 10.9 GB | |
| 4 | Gemma 3 4B ITGoogle Q4_K_M · 5.0 GB · 209 t/s | 209 tok/s | Q4_K_M · 5.0 GB | |
| 5 | Q4_K_M · 6.5 GB · 95.5 t/s | 95.5 tok/s | Q4_K_M · 6.5 GB |
The RTX 3080 Ti is the quickest card in the 12 GB class: it runs 7–9B models well. It is among the most efficient per watt in its tier, and at its $1,199 MSRP it cost more per token than most of the cards compared. Ampere generation, released 2021.
RTX 3080 Ti strengths and weaknesses
Strengths
- Runs 14B models: Qwen3 14B at 62.4 tok/s in 11.9 GB
Weaknesses
- 27–31B models don't fit: Qwen3.8 27B needs 17.9 GB at Q4 vs 12 GB here
- Little room for context: Qwen3 14B uses 11.9 of 12 GB, 0.1 GB left
- No FP8 or NVFP4: Ampere predates both; use GGUF, AWQ or GPTQ builds
Speed per dollar and per watt
On Qwen3.5 9B (9B params) at Q4_K_M. New prices are compared with new, used with used. The last two columns are the speed you get per $100 of price and per 100 W of board power; higher is better. The RTX 3080 Ti has no current price, so it is shown at its MSRP.
| Card | New price | tok/s | tok/s per $100 | tok/s per 100 W |
|---|---|---|---|---|
RTX 50608 GBBest new value | about $460 | 45.8 | 10.0 | 31.6 |
RTX 507012 GB | about $840 | 68.7 | 8.2 | 27.5 |
RTX 5070 Ti16 GB | about $1,170 | 91.6 | 7.8 | 30.5 |
RTX 3080 Ti12 GB | about $1,199 | 93.2 | 7.8 | 26.6 |
RX 907016 GB | about $800 | 60.4 | 7.6 | 27.5 |
RTX 508016 GB | about $1,690 | 98.1 | 5.8 | 27.3 |
RX 9060 XT 16GB16 GB | about $590 | 30.2 | 5.1 | 18.9 |
RTX 509032 GB | about $5,376 | 183 | 3.4 | 31.8 |
| Card | Used price | tok/s | tok/s per $100 | tok/s per 100 W |
|---|---|---|---|---|
RTX 50608 GBBest used value | about $388 | 45.8 | 11.8 | 31.6 |
RTX 5070 Ti16 GB | about $1,100 | 91.6 | 8.3 | 30.5 |
RTX 408016 GB | about $1,100 | 73.3 | 6.7 | 22.9 |
RTX 4080 SUPER16 GB | about $1,155 | 75.2 | 6.5 | 23.5 |
RTX 3090 Ti24 GB | about $1,600 | 103 | 6.4 | 22.9 |
RX 9060 XT 16GB16 GB | about $480 | 30.2 | 6.3 | 18.9 |
RTX 508016 GB | about $1,640 | 98.1 | 6.0 | 27.3 |
RTX 509032 GB | about $6,400 | 183 | 2.9 | 31.8 |
Cost per token: the RTX 3080 Ti against 8–32 GB cards
On Qwen3.5 9B (9B params, Q4_K_M, 4K context). Speed, cost per token and power per token depend on the model.
- Card, MSRP
- $0.82
- Power
- $0.16
Renting costs as much as buying after ≈9,800 hours (≈6.7 years at 4 h a day).
Each dot is a card; hover or tap it for its numbers, click to open it. Cards without a price for the chosen basis are left out. How we calculate cost per token
Cards people weigh against the RTX 3080 Ti, each with the one reason it is worth a look. Speeds on Qwen3.5 9B.
Output speed depends on memory bandwidth (912 GB/s). Grey rows need more than 12 GB.
- 9BQwen3.5 9B · Q4_K_M · 6.6 GB93.2tok/s
- 62.4tok/s
- 27–31BQwen3.8 27B · not at 4-bitneeds 17.9 GB
- 35B MoEQwen3.6 35B A3B · not at 4-bitneeds 22.0 GB
- 70BLlama 3.3 70B Instruct · not at 4-bitneeds 51.3 GB
- 120B MoEGPT OSS 120B · not at 4-bitneeds 70.8 GB
Estimated speeds: Q4_K_M, 4K context, one user. How we estimate
All 956 models on the RTX 3080 Ti
Featured onlyBest quant that fits, ordered by llmrun Score. Fit grades how well it runs here, S best. How ranking works
| Model | Quant | VRAM | Speed | Context | Score | Fit |
|---|---|---|---|---|---|---|
Crowd V1ChatFunctions | BF16 | 3.7 GB | 267.1 t/s | — | S100, Runs great | |
| Q4_K_M | 1.5 GB | 502.6 t/s | 1049K | S100, Runs great | ||
Qwen3 VL 8B InstructVisionFunctions | Q4_K_M | 8.0 GB | 101.2 t/s | 262K | S92, Runs great | |
| Q4_K_M | 2.6 GB | 599.1 t/s | — | S100, Runs great | ||
Qwen3 4B Mamba3 Hybrid Stage2b Kd Bias0ChatFunctions | Q4_K_M | 4.7 GB | 178.1 t/s | 41K | S99, Runs great | |
JiRackUltra 7BChatFunctions | Q4_K_M | 5.8 GB | 118.8 t/s | 131K | S95, Runs great | |
SmolLM3 3BChatFunctions | Q4_K_M | 3.4 GB | 257.9 t/s | 66K | S100, Runs great | |
Wald 4BChatReasoningFunctions | Q4_K_M | 3.4 GB | 181.9 t/s | 262K | S99, Runs great | |
Molmo2 4BVision | Q4_K_M | 4.7 GB | 174.4 t/s | 37K | S98, Runs great | |
Gemma4 12B QAT Uncensored HauhauCS BalancedVisionRoleplayChat | Q4_K_M | 10.8 GB | 74.9 t/s | — | A72, Runs well | |
Sevclassifier Jev Style 1.5BChatFunctions | Q4_K_M | 1.7 GB | 463.3 t/s | 33K | S100, Runs great | |
| Q4_K_M | 0.4 GB | 1520.7 t/s | 2K | S100, Runs great | ||
CAJAL 4BChat | Q4_K_M | 4.8 GB | 224.6 t/s | — | S100, Runs great | |
Ternary Bonsai 8B UnpackedChatFunctions | Q4_K_M | 7.6 GB | 107.4 t/s | 66K | S92, Runs great | |
Hy MT2 7BTranslation | Q4_K_M | 7.3 GB | 110.0 t/s | 262K | S93, Runs great |
Specs
- Architecture
- Ampere
- VRAM
- 12 GB GDDR6X
- Memory bandwidth
- 912.4 GB/s
- FP16 (dense)
- 68.20 TFLOPS
- CUDA cores
- 10,240
- Tensor cores
- 320
- Compute capability
- 8.6
- TDP
- 350 W
- Released
- 2021-06-03
- MSRP
- $1,199
RTX 3080 Ti FAQ
Is the RTX 3080 Ti still worth it in 2026?
At its MSRP of $1,199 you get 7.8 tok/s per $100 on Qwen3.5 9B, which ranks 4th of the 8 cards compared here; it is also the fastest out of 9 cards with 12 GB. It runs 7–9B models well; 14B only just fits.
What is the largest model the RTX 3080 Ti can run?
GPT OSS 20B (20.9B) is the largest at 4-bit or better: at IQ4_XS it uses 11.8 GB of 12 GB, and runs at about 191 tok/s.
Can the RTX 3080 Ti run Qwen3.5 9B?
Yes. Qwen3.5 9B fits at Q4_K_M, using 6.6 GB of 12 GB, and it is the best model that runs well on this card.
Can the RTX 3080 Ti run 27–31B models?
No. Qwen3.8 27B needs 17.9 GB at Q4, more than the 12 GB on this card.
How many tokens per second does the RTX 3080 Ti give?
About 93.2 tok/s on Qwen3.5 9B at Q4_K_M, with the first token after 2.6 s on a 4K-token prompt.
What power supply does the RTX 3080 Ti need?
Plan a 700 W PSU (2× the 350 W TDP).
Is 12 GB enough for coding and agents?
Yes for coding: Llama 3.1 8B Instruct, the only coding-scored model that fits, runs at about 112 tok/s. For agents, not yet: no agent-scored model fits in less than 96 GB.
Is renting cheaper than buying an RTX 3080 Ti?
Rent from $0.12/hr. Renting costs as much as buying after ≈9,800 hours (≈6.7 years at 4 h a day).