NVIDIA GeForce GTX 1650
4 GB decides what fits. 128.1 GB/s decides how fast it runs once it does.
Open in the calculator →Best models for 4 GB, on every card that size →
Find one: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
- Largest popular model that fitsQwen3.5-4B4.7B params · Q4_K_M · needs 3.98 GB~32 tok/s Faster than you readOpen in the calculator →
At 8,192 tokens of context with the whole model in device memory. Speeds are estimates from memory bandwidth, not benchmarks run on this card.
70 of 320 fit entirely
70 of the 320 models the engine can size fit entirely in device memory at Q4_K_M where it is published, otherwise the nearest published format, and 8,192 tokens. 141 run with some layers on system memory (32 GB assumed), and 109 do not run at all.
Featured models · Q4_K_M at 8,192 tokens, including offload
| Model | Needs | Verdict | Decode | Calculator |
|---|---|---|---|---|
| Qwen3.5-2B2.3B parameters | 2.32 GB | fits | ~71 tok/s | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF164.0B parameters | 3.46 GB | fits | ~33 tok/s | Open → |
| Qwen3-VL-8B-Instruct8.8B parameters | 7.04 GB | 22 of 36 layers on system RAM (3.07 GB) | ~12 tok/s with offload | Open → |
| Qwen3.5-9B9.7B parameters | 7.05 GB | 17 of 32 layers on system RAM (3.15 GB) | ~15 tok/s with offload | Open → |
| gemma-4-12B-it12B parameters | 9.54 GB | 34 of 48 layers on system RAM (5.57 GB) | ~8.3 tok/s with offload | Open → |
| gpt-oss-20b21B parameters | 13.8 GB | 19 of 24 layers on system RAM (10.1 GB) | ~26 tok/s with offload | Open → |
| Qwen3.8-27B28B parameters | 18.6 GB | 55 of 64 layers on system RAM (14.7 GB) | ~4.3 tok/s with offload | Open → |
| Qwen3.6-27B28B parameters | 18.6 GB | 55 of 64 layers on system RAM (14.7 GB) | ~4.3 tok/s with offload | Open → |
| Kimi-Linear-48B-A3B-Instruct49B parameters | 30.4 GB | 25 of 27 layers on system RAM (27.3 GB) | ~33 tok/s with offload | Open → |
| Qwen3.8-Flash-Next180B parameters | 111.6 GB | needs 108.2 GB of system RAM; 32 GB assumed | does not run | Open → |
| DeepSeek-V4-Flash-Vision-Exp305B parameters | 188.1 GB | needs 187.2 GB of system RAM; 32 GB assumed | does not run | Open → |
| GLM-5.3-Flash321B parameters | 198.3 GB | needs 197.3 GB of system RAM; 32 GB assumed | does not run | Open → |
These examples are selected from prominent labs using the catalogue’s latest Hugging Face 30-day downloads and repository-creation freshness signal, with newer releases guaranteed a place. Offloaded rows assume 32 GB of system RAM, and a speed is only shown for a row that runs. Fitting in memory is not the same as loading: whether the runtime and version you have supports each architecture and format on this machine has not been tested here. Each “Open” link carries the same model, format, context and RAM into the calculator.
Fits entirely in NVIDIA GeForce GTX 1650 memory — 70 of 320 sized models
The most demanding model that fits is Qwen3.5-4B at Q4_K_M: 3.98 GB of the 4 GB, leaving 0.02 GB spare.
Fully resident at 8,192 tokens (or the model’s own maximum, where that is shorter), offload off, at 128.1 GB/s. Each row is a run of the engine for this configuration; the rows start with current, prominent releases and “fits” is memory, not a tested runtime. Older or less prominent models remain available through this search and “Show all”.
Showing 40 of 70 models that fit.
| Model | Format | Needs | Spare | Decode | Downloads / 30d | Calculator |
|---|---|---|---|---|---|---|
| Qwen3.5-4BQwen · 4.7B params | Q4_K_M | 3.98 GB | 0.02 GB | ~32 tok/s | 7.8M | Open → |
| Qwen3.5-2BQwen · 2.3B params | Q4_K_M | 2.32 GB | 1.68 GB | ~71 tok/s | 4.9M | Open → |
| NVIDIA-Nemotron-3-Nano-4B-BF16NVIDIA · 4.0B params | Q4_K_M | 3.46 GB | 0.54 GB | ~33 tok/s | 3.4M | Open → |
| Qwen3.5-0.8BQwen · 873M params | Q4_K_M | 1.46 GB | 2.54 GB | ~155 tok/s | 2.6M | Open → |
| North-Micro-Vision-InstructCohereLabs · 2.5B params | Q4_K_M | 2.99 GB | 1.01 GB | ~46 tok/s | 180.7K | Open → |
| granite-4.1-3bIBM · 3.4B params | Q4_K_M | 3.72 GB | 0.28 GB | ~31 tok/s | 521.8K | Open → |
| LFM2.5-2.6BLiquidAI · 2.7B params | Q4_K_M | 2.61 GB | 1.39 GB | ~50 tok/s | 108.1K | Open → |
| Qwen3-VL-2B-InstructQwen · 2.1B params | Q4_K_M | 3.05 GB | 0.95 GB | ~43 tok/s | 2.8M | Open → |
| granite-4.2-3bIBM · 3.7B params | Q4_K_M | 3.72 GB | 0.28 GB | ~31 tok/s | 50.6K | Open → |
| LFM2.5-230MLiquidAI · 230M params | Q4_K_M | 1.05 GB | 2.95 GB | ~352 tok/s | 88.6K | Open → |
| Qwen3-0.6BQwen · 752M params | Q4_K_M | 2.22 GB | 1.78 GB | ~65 tok/s | 29.7M | Open → |
| LFM2.5-VL-3BLiquidAI · 3.1B params | Q4_K_M | 2.85 GB | 1.15 GB | ~50 tok/s | 26.7K | Open → |
| LFM2.5-350MLiquidAI · 354M params | Q4_K_M | 1.13 GB | 2.87 GB | ~227 tok/s | 69.9K | Open → |
| Hy-MT2-1.8BTencent · 2.0B params | Q4_K_M | 2.47 GB | 1.53 GB | ~52 tok/s | 28.9K | Open → |
| LFM2.5-1.2B-InstructLiquidAI · 1.2B params | Q4_K_M | 1.63 GB | 2.37 GB | ~84 tok/s | 119.2K | Open → |
| Qwen3-1.7BQwen · 2.0B params | Q4_K_M | 3.02 GB | 0.98 GB | ~43 tok/s | 3.1M | Open → |
| Nemotron-3.5-Content-SafetyNVIDIA · 4.3B params | Q4_K_M | 3.73 GB | 0.27 GB | ~31 tok/s | 14.1K | Open → |
| SmolLM3-3BHuggingFaceTB · 3.1B params | Q4_K_M | 3.46 GB | 0.54 GB | ~35 tok/s | 617K | Open → |
| Ministral-3-3B-Instruct-2512Mistral AI · 3.8B params | Q4_K_M | 3.82 GB | 0.18 GB | ~29 tok/sAbout reading pace | 114.1K | Open → |
| Qwen3-1.7B-BaseQwen · 1.7B params | Q4_K_M | 3.02 GB | 0.98 GB | ~43 tok/s | 696K | Open → |
| Qwen2.5-VL-3B-InstructQwen · 3.8B params | Q4_K_M | 3.41 GB | 0.59 GB | ~39 tok/s | 2.4M | Open → |
| Qwen2.5-0.5B-InstructQwen · 494M params | Q4_K_M | 1.31 GB | 2.69 GB | ~203 tok/s | 8.7M | Open → |
| Qwen2.5-1.5B-InstructQwen · 1.5B params | Q4_K_M | 2.15 GB | 1.85 GB | ~72 tok/s | 7.4M | Open → |
| OLMo-2-0425-1Ballenai · 1.5B params · sized at 4,096 tokens, its maximum | Q4_K_M | 2.27 GB | 1.73 GB | ~64 tok/s | 404.2K | Open → |
| SmolLM3-3B-BaseHuggingFaceTB · 3.1B params | Q4_K_M | 3.46 GB | 0.54 GB | ~35 tok/s | 133.5K | Open → |
| DeepSeek-R1-Distill-Qwen-1.5BDeepSeek · 1.8B params | Q4_K_M | 2.15 GB | 1.85 GB | ~72 tok/s | 961.3K | Open → |
| Qwen2.5-3B-InstructQwen · 3.1B params | Q4_K_M | 3.21 GB | 0.79 GB | ~39 tok/s | 4.2M | Open → |
| SmolLM2-135M-InstructHuggingFaceTB · 135M params | Q4_K_M | 1.09 GB | 2.91 GB | ~316 tok/s | 1.8M | Open → |
| Phi-tiny-MoE-instructMicrosoft · 3.8B params · sized at 4,096 tokens, its maximum | Q4_K_M | 3.22 GB | 0.78 GB | ~102 tok/s | 71.7K | Open → |
| SmolLM2-135MHuggingFaceTB · 135M params | Q4_K_M | 1.09 GB | 2.91 GB | ~316 tok/s | 1.8M | Open → |
| Qwen2.5-0.5BQwen · 494M params | Q4_K_M | 1.31 GB | 2.69 GB | ~203 tok/s | 1.5M | Open → |
| gemma-3-4b-itGoogle · file from the unsloth mirror · 4.3B params | Q4_K_M | 3.58 GB | 0.42 GB | ~33 tok/s | 98.8K | Open → |
| Qwen2.5-Coder-3B-InstructQwen · 3.1B params | Q4_K_M | 3.21 GB | 0.79 GB | ~39 tok/s | 532.9K | Open → |
| granite-4.0-h-microIBM · 3.2B params | Q4_K_M | 2.89 GB | 1.11 GB | ~42 tok/s | 6.8K | Open → |
| phi-2Microsoft · 2.8B params · sized at 2,048 tokens, its maximum | Q4_K_M | 3.18 GB | 0.82 GB | ~31 tok/s | 599.2K | Open → |
| SmolLM2-360MHuggingFaceTB · 362M params | Q4_K_M | 1.39 GB | 2.61 GB | ~155 tok/s | 411.5K | Open → |
| PowerMoE-3bIBM · 3.4B params · sized at 4,096 tokens, its maximum | Q4_K_M | 3.13 GB | 0.87 GB | ~106 tok/s | 552.6K | Open → |
| gemma-3-1b-itGoogle · file from the unsloth mirror · 1000M params | Q4_K_M | 1.66 GB | 2.34 GB | ~128 tok/s | 59.8K | Open → |
| Qwen2.5-Coder-1.5B-InstructQwen · 1.5B params | Q4_K_M | 2.15 GB | 1.85 GB | ~72 tok/s | 497.3K | Open → |
| Llama-3.2-1B-InstructMeta · file from the unsloth mirror · 1.2B params | Q4_K_M | 2.02 GB | 1.98 GB | ~81 tok/s | 488.5K | Open → |
Sized at the format most people actually download, not at FP16. The catalogue ordered by downloads →
Find the NVIDIA GeForce GTX 1650: Amazon ↗ · eBay (new and used) ↗Store links may pay us a commission. They never decide which card is suggested — the memory arithmetic does.
The GTX 1650, model by model
One page per model: whether it fits this device, at which formats, and how fast.
- Qwen3-0.6B
- Qwen3-VL-8B-Instruct
- gemma-4-26B-A4B-it
- Qwen3-8B
- gemma-4-31B-it
- Qwen3.5-9B
- Qwen2.5-0.5B-Instruct
- Qwen2.5-7B-Instruct
- Qwen3.5-4B
- Qwen3-4B
- Qwen2.5-1.5B-Instruct
- Qwen3.8-27B
All 70Fewer
- gpt-oss-20b
- Qwen2.5-VL-7B-Instruct
- GLM-5.3-Flash
- Qwen3.5-2B
- dolphin-2.9.1-yi-1.5-34b
- DeepSeek-V4-Flash-0731
- gpt-oss-120b
- gemma-4-E4B-it
- Qwen2.5-3B-Instruct
- Qwen3-32B
- Qwen3-4B-Instruct-2507
- DeepSeek-V3.2
- Qwen3-VL-4B-Instruct
- pythia-160m
- NVIDIA-Nemotron-3-Nano-4B-BF16
- Qwen3.6-35B-A3B
- Qwen3-1.7B
- gemma-4-E2B-it
- OTel-2.0-LLM-31B-IT
- Qwen3-VL-2B-Instruct
- Qwen3-14B
- Qwen3.5-0.8B
- JiRackUltra_1b
- Qwen3.6-27B
- Qwen2.5-VL-3B-Instruct
- Qwen2.5-Coder-7B-Instruct
- Mistral-7B-Instruct-v0.3
- Qwen2.5-32B-Instruct
- gemma-4-12B-it
- Qwen3.5-27B
- SmolLM2-135M-Instruct
- SmolLM2-135M
- GLM-4.7-Flash
- Mistral-7B-Instruct-v0.2
- Qwen2.5-Coder-14B-Instruct
- Qwen2.5-14B-Instruct
- Qwen3.5-35B-A3B
- Qwen3-30B-A3B
- Qwen2.5-0.5B
- Ornith-1.0-35B
- pythia-70m-deduped
- DeepSeek-V3-0324
- Qwen3.8-Flash-Next
- GLM-5.3
- TinyLlama-1.1B-Chat-v1.0
- DeepSeek-V3
- Kimi-K3
- NVIDIA-Nemotron-3-Super-120B-A12B-BF16
- MiniMax-M2.7
- DeepSeek-V4-Flash
- MiniCPM5-2B
- DeepSeek-R1
- DeepSeek-R1-Distill-Qwen-1.5B
- Ornith-1.0-9B
- DeepSeek-V4-Flash-Vision-Exp
- DeepSeek-Coder-V2-Lite-Instruct
- Qwen2.5-Coder-32B-Instruct
- NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
The specification behind every figure
What the manufacturer publishes for this device, and the pages it was read from.
Manufacturer specification
| Memory scope | dedicated |
|---|---|
| Capacity | 4 GB |
| Published options | 4 GB |
| Bandwidth | 128.1 GB/s |
| Memory type | GDDR5 |
| Bus width | 128-bit |
| FP32 peak | Not published |
| Dense matrix peak | Not published without sparsity |
| Graphics Card Power | 75 W |
Source ledger
Caveats
- NVIDIA does not publish this card's memory bandwidth any more, so the figure comes from the TechPowerUp GPU Database (https://www.techpowerup.com/gpu-specs/geforce-gtx-1650.c3366), a third-party source, and is labelled as an estimate rather than a published specification.
- No dense matrix throughput is published for this device in a form this catalogue accepts, so no compute roof is priced and time to first token is withheld.
Published capacity is a hardware ceiling, not guaranteed free runtime memory. The calculator shows the runtime reserve separately rather than folding it into a single number.
Run the diagnostic on the NVIDIA GeForce GTX 1650 →