Calculator Models Hardware Compare

Product

Calculator
Compare
Tier List

Browse

Models
Hardware
Docs

About

Why It Works
What's New
Legal Notice
Privacy Policy

All estimates are approximations based on mathematical models and public specifications. Actual performance may vary. Do not make purchasing decisions based solely on these estimates.

Data sourced from Hugging Face, Ollama, and official model documentation. Model names and logos are trademarks of their respective owners.

© 2026 Will It Run AI — Fase Consulting Ibiza, S.L. (NIF: B57969656)

Home/Hardware/GPUs/GTX 1650 4GB

NVIDIA

GTX 1650 4GB

GTX 16ConsumerTuringPCIe 3CUDA

4GB

VRAM

128GB/s

Bandwidth

6TFLOPS

FP16 Compute

24TOPS

INT8 Inference

GTX 1650 4GBCategory AvgIntel Arc A380 6GB

Specifications

Compute

FP166 TFLOPS

INT824 TOPS

ArchitectureTuring

Memory

VRAM4 GB

Bandwidth128 GB/s

General

FamilyGTX 16

SegmentConsumer

InterconnectPCIe 3

Compute PlatformCUDA

Architecture

Turing

Turing is NVIDIA's first-generation RTX architecture, introducing dedicated RT and Tensor Cores to consumer GPUs for the first time. Built on TSMC's 12nm FinFET process.

AI Relevance

The first consumer architecture with Tensor Cores, enabling meaningful acceleration for INT8 and FP16 inference. However, limited VRAM (typically 6-11 GB) restricts modern LLM model sizes.

Process: TSMC 12nmPlatform: CUDATensor Cores: Gen 2Precisions: FP32, FP16, INT8, INT4

Recommendations by Workload

Agentic Coding

C

Qwen 2.5 Coder 1.5B

This model is still usable for agentic-coding, but it is not the most specialized pick. It sits in the middle of the current model mix. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.

Decode 64.0 tok/s · 33K ctx · llama.cpp

3.0 GB / 4.0 GB VRAM

Chat

C

Qwen 3 1.7B

This model is a direct match for chat. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.

Decode 61.7 tok/s · 10K ctx · llama.cpp

3.1 GB / 4.0 GB VRAM

Coding

C

Qwen 2.5 Coder 1.5B

This model is still usable for coding, but it is not the most specialized pick. It sits in the middle of the current model mix. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.

Decode 64.0 tok/s · 21K ctx · llama.cpp

3.0 GB / 4.0 GB VRAM

RAG

C

Qwen 3 1.7B

This model is still usable for rag, but it is not the most specialized pick. It belongs to a current frontier family for local AI. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.

Decode 61.7 tok/s · 33K ctx · llama.cpp

3.1 GB / 4.0 GB VRAM

Reasoning

C

DeepSeek R1 1.5B

This model is a direct match for reasoning. It sits in the middle of the current model mix. It fits natively with comfortable headroom. Known channels: huggingface, ollama, lm-studio.

Decode 64.0 tok/s · 21K ctx · llama.cpp

3.0 GB / 4.0 GB VRAM

Full Model Compatibility

Llama 3.2 1B Instruct Q8 0

1B2.9 GB67 tok/s22K ctx

Qwen2.5 1.5B Instruct

1.5B3.0 GB64 tok/s21K ctx

TinyLlama 1.1B Chat v1.0

1.1B2.8 GB64 tok/s23K ctx

SmolVLM 500M Instruct

0.5B2.5 GB67 tok/s25K ctx

embeddinggemma 300M

0.3B2.3 GB67 tok/s27K ctx

2B3.3 GB53 tok/s19K ctx

Gemmasutra Mini 2B v1

2B3.3 GB53 tok/s19K ctx

2B3.7 GB41 tok/s17K ctx

Qwen2.5 3B Instruct

3B3.9 GB35 tok/s16K ctx

Llama 3.2 3B Instruct

3B4.3 GB29 tok/s15K ctx

4B4.5 GB23 tok/s14K ctx

4B4.5 GB23 tok/s14K ctx

DeepSeek R1 671B

671B416.4 GB2 tok/s4K ctx

Devstral 2 123B Instruct

123B95.5 GB2 tok/s4K ctx

744B461.4 GB2 tok/s4K ctx

27B22.0 GB4 tok/s4K ctx

Qwen3.5 35B A3B

35B28.1 GB3 tok/s4K ctx

9B8.2 GB12 tok/s8K ctx

1000B616.3 GB2 tok/s4K ctx

Mistral Large 3

675B419.5 GB2 tok/s4K ctx

Mistral Small 4 119B

119B74.9 GB3 tok/s4K ctx

Qwen3-Coder 30B A3B Instruct

30.5B20.7 GB9 tok/s4K ctx

Qwen3-Coder 480B A35B Instruct

480B299.6 GB2 tok/s4K ctx

Qwen3-Coder-Next

80B50.9 GB4 tok/s4K ctx

Qwen3.5 9B Uncensored HauhauCS Aggressive

9B8.2 GB12 tok/s8K ctx

Qwen3.5 122B A10B

122B80.1 GB2 tok/s4K ctx

Meta Llama 3.1 8B Instruct

8B7.4 GB13 tok/s9K ctx

DeepSeek V3 671B

671B416.4 GB2 tok/s4K ctx

141B93.4 GB2 tok/s4K ctx

72B56.5 GB2 tok/s4K ctx

Qwen 3 235B A22B

235B148.1 GB2 tok/s4K ctx

Qwen3-VL 30B A3B Instruct

30B20.4 GB9 tok/s4K ctx

Llama 2 7B Chat

7B6.7 GB15 tok/s10K ctx

llava llama 3 8b v1 1

8B7.4 GB13 tok/s9K ctx

Qwen3.5 397B A17B

397B305.5 GB2 tok/s4K ctx

Devstral Small 2 24B Instruct

24B19.7 GB4 tok/s4K ctx

70B54.9 GB2 tok/s4K ctx

Llama 4 Maverick 17B 128E

400B248.0 GB2 tok/s4K ctx

Mistral 7B Instruct v0.2

7B6.7 GB15 tok/s10K ctx

DeepSeek R1 0528 Qwen3 8B

8B7.4 GB13 tok/s9K ctx

111B86.4 GB2 tok/s4K ctx

Qwen 2.5 Coder 32B

32B25.8 GB3 tok/s4K ctx

Qwen 2.5 VL 72B

72B56.5 GB2 tok/s4K ctx

27B22.0 GB4 tok/s4K ctx

9B8.2 GB12 tok/s8K ctx

Mistral 7B Instruct v0.3

7B6.7 GB15 tok/s10K ctx

Meta Llama 3 8B Instruct

8B7.4 GB13 tok/s9K ctx

Qwen3.5 35B A3B

35B28.1 GB3 tok/s4K ctx

Codestral 2 25.08

22B18.2 GB5 tok/s4K ctx

Devstral Small 1.1

24B19.7 GB4 tok/s4K ctx

Just out of reach

Models you could run with an upgrade

High-quality models that need a bit more memory

DeepSeek R1 671B

671BTier 5Needs ~422.2 GB

Devstral 2 123B Instruct

123BTier 5Needs ~114.8 GB

Runs on Mac Studio M3 Ultra 256GB

744BTier 5Needs ~467.6 GB

27BTier 5Needs ~26.2 GB

Runs on RTX 5090 32GB (~$1,999)

Qwen3.5 35B A3B

35BTier 5Needs ~33.6 GB

Runs on Mac mini M4 64GB (~$1,099)

Upgrade paths

Upgrade from GTX 1650 4GB

See what you unlock with more powerful hardware

Upgrade options

Upgrade options

Intel Arc A380 6GBNext step up

6 GB VRAM (+2)186 GB/s (+58)

Unlocks Qwen3.5 4B, gemma 3 4b it, gemma 3 4b it+14 more · +24% faster avg

GTX 1660 Super 6GBNVIDIA upgrade

6 GB VRAM (+2)336 GB/s (+208)

Unlocks Qwen3.5 4B, gemma 3 4b it, gemma 3 4b it+14 more · +151% faster avg

Intel Arc B580 12GBBest value

12 GB VRAM (+8)456 GB/s (+328)

Unlocks Qwen3.5 9B, Qwen3.5 9B Uncensored HauhauCS Aggressive, Meta Llama 3.1 8B Instruct+150 more · +62% faster avg

~$249 MSRP

AMD Instinct MI350X 288GBBiggest leap

288 GB VRAM (+284)8000 GB/s (+7872)

Unlocks Devstral 2 123B Instruct, Qwen3.5 27B, Qwen3.5 35B A3B+260 more · +3001% faster avg

~$8,000 MSRP

Compare this GPU