InternVL2 2B — Hardware Requirements & GPU Compatibility
Vision
InternVL2-2B is OpenGVLab's 2.2-billion-parameter instruction-tuned vision-language model, part of the InternVL 2.0 family that ranges from 1 billion to 108 billion parameters. It pairs the InternViT-300M-448px vision encoder with the internlm2-chat-1.8b language model through an MLP projector, and is trained to handle document and chart comprehension, infographics QA, scene-text and OCR tasks, scientific and mathematical problem solving, and multi-image or video input. Compared with earlier Mini-InternVL models it adds support for long texts, multiple images, and video within the same training context. At just over 2 billion parameters, it runs comfortably on a single consumer GPU or even weaker hardware. Context length is 8,192 tokens, matching its 8k training context window. It is released under the MIT license, permitting unrestricted commercial and research use. It was published in June 2024, and has since been superseded by the InternVL2.5 and InternVL3 series.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 32K)Best on 24 GB (Q4_K_M · ~449 tok/s · 32K)Macs from 8 GB (Q4_K_M · ~49 tok/s)
MIT
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run InternVL2 2B
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
vllm serve OpenGVLab/InternVL2-2B --max-model-len 16384vLLM repo with its own quantisation
Download: Weights on Hugging Face
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (8 GB headroom) · Q4_K_M
- Generation speed
- ~406 tok/s
- generation speed
- Cost per 1M output tokens
- $0.084
- per 1M output tokens
How Much VRAM Does InternVL2 2B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 1.0 GB | 4.9 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 32K context · ~636 tok/s | |
| 1.2 GB | 5 GB | 3-bit medium — 3-bit medium quantization | Fits: 32K context · ~555 tok/s | |
| 1.5 GB | 5.3 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 32K context · ~449 tok/s | |
| 1.7 GB | 5.6 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 32K context · ~379 tok/s | |
| 2 GB | 5.8 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 32K context · ~328 tok/s | |
| 2.4 GB | 6.3 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 32K context · ~270 tok/s | |
| 4.8 GB | 8.7 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 32K context · ~135 tok/s |
Selected Q4_K_M
4-bit medium quantization3.3GBat 16K context
- Weights
- 0.9 GB
- KV cache
- 2.1 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 20.8 GB free · ~449 tok/s
≥50 tok/s needs ≥112 GB/s · RTX 5090 ~798, RTX 3090 Ti ~449
Which Hardware Can Run InternVL2 2B?
Q4_K_M · 1.5 GBInternVL2 2B (Q4_K_M) requires 1.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 2+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti. 163 devices with unified memory can run InternVL2 2B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB).
≤ 11 GB
≤ 11 GB tier →fits · 4.8 GB free · up to 32K ctx
16 GPUs · 1 Mac class · 2 device families
RTX 3080~339Simulate output on RTX 3080RTX 2080 Ti~274Simulate output on RTX 2080 TiRTX 3070 Ti~271Simulate output on RTX 3070 TiM3*8 GB~49Simulate output on iMac 24" M3 (8 GB)Orin Nano~45Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~33Simulate output on iPhone 1720 GB
fits · 16.8 GB free · up to 32K ctx
1 GPU
RX 7900 XT~329Simulate output on RX 7900 XT32–40 GB
fits · 28.8 GB free · up to 32K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~798Simulate output on RTX 5090M5 Max36 GB~221Simulate output on Mac Studio M5 Max (36 GB)M532 GB~74Simulate output on MacBook Air 13" M5 (32 GB)Orin~91Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~60Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| LFM3 Omni 2BLiquid AI | — | 2.4 GB | ~385 | same size |
| Qwen3.5 2BAlibaba | — | 1.9 GB | ~370 | same size |
| Qwen3 VL 2B InstructAlibaba | — | 3.5 GB | ~362 | same size |
| Qwen2 VL 2B InstructAlibaba | — | 2.1 GB | ~390 | same size |
| Kimi K3 DSparkRadixArk | — | 2.2 GB | ~381 | same size |
| Qwen3.8 2Bempero-ai | — | 1.9 GB | ~370 | same size |
Specifications
- Publisher
- OpenGVLab
- Parameters
- 2.2B
- Architecture
- InternVLChatModel
- Release Date
- 2024-06-27
- License
- MIT
Frequently Asked Questions
How much VRAM does InternVL2 2B need?
InternVL2 2B requires 1.5 GB of VRAM at Q4_K_M, or 4.8 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 2.2B × 4.8 bits ÷ 8 = 1.3 GB
KV Cache + Overhead ≈ 0.2 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
VRAM usage by quantization
What's the best quantization for InternVL2 2B?
For InternVL2 2B, Q4_K_M (1.5 GB) offers the best balance of quality and VRAM usage. Q5_K_M (1.7 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 1.0 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run InternVL2 2B on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 145 other Macs can run InternVL2 2B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 1.5 GB of usable unified memory (RAM minus macOS overhead).
Can I run InternVL2 2B locally?
Yes — InternVL2 2B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is InternVL2 2B?
At Q4_K_M, InternVL2 2B can reach ~3288 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~449 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 1.46 × 0.65 = ~3562 tok/s
Estimated speed at Q4_K_M (1.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of InternVL2 2B?
At Q4_K_M, the download is about 1.32 GB. The full-precision BF16 version is 4.41 GB. The smallest option (Q2_K) is 0.94 GB.
Which GPUs can run InternVL2 2B?
52 consumer GPUs can run InternVL2 2B at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT. 52 GPUs have plenty of headroom for comfortable inference.
Which devices can run InternVL2 2B?
163 devices with unified memory can run InternVL2 2B at Q4_K_M (1.5 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.