Intern Decision 4B — Hardware Requirements & GPU Compatibility
VisionFunctionsIntern-Decision-4B is InternLM's 4.5-billion-parameter multimodal structured decision model, fine-tuned from Qwen3.5-4B. It accepts a shared state, a schema of named questions and optional images, and returns an answer distribution for every question in one forward pass by scoring candidate option symbols instead of sampling text. The card reports an average of 90.02 across its decision benchmarks, the highest in its family table, with roughly 44 ms per query on a single RTX 4090. At this size it fits on a single consumer GPU, including mid-range cards. The model supports a 262,144 token context window. It is released under the Apache 2.0 license, permitting unrestricted commercial and research use, and was published in September 2026. It is the largest of the Intern-Decision models listed in the card, alongside 0.8B and 2B siblings, and it depends on the custom inference code in the repository rather than a standard chat workflow.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 128K)Best on 24 GB (Q4_K_M · ~205 tok/s · 256K)Macs from 8 GB (Q4_K_M · ~23 tok/s)
Apache 2.0
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Intern Decision 4B
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
ollama run hf.co/yeaay/Intern-Decision-4B-GGUF:Q4_K_MQ4_K_M fits fully on the RTX 4090 · 20.6 GB free · 16K context
Download: Intern-Decision-4B-Q4_K_M.gguf2.71 GByeaay
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3060 12GB (8 GB headroom) · Q4_K_M
- Generation speed
- ~73 tok/s
- generation speed
- Cost per 1M output tokens
- $0.24
- per 1M output tokens
How Much VRAM Does Intern Decision 4B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 2.4 GB | 2.9 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 256K context · ~273 tok/s | |
| 2.7 GB | 3.2 GB | 3-bit medium — 3-bit medium quantization | Fits: 256K context · ~245 tok/s | |
| 3.2 GB | 3.7 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 256K context · ~205 tok/s | |
| 3.7 GB | 4.2 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 256K context · ~177 tok/s | |
| 4.2 GB | 4.7 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 256K context · ~156 tok/s | |
| 5.0 GB | 5.5 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 256K context · ~131 tok/s | |
| 9.6 GB | 10.1 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Fits: 256K context · ~69 tok/s |
Selected Q4_K_M
4-bit medium quantization3.4GBat 16K context
- Weights
- 2.7 GB
- KV cache
- 0.3 GB
- Overhead
- 0.3 GB
Which Hardware Can Run Intern Decision 4B?
Q4_K_M · 3.2 GBIntern Decision 4B (Q4_K_M) requires 3.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 5+ GB is recommended. Using the full 262K context window can add up to 5.2 GB, bringing total usage to 8.4 GB. 114 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce GTX 1650 SUPER. 185 devices with unified memory can run Intern Decision 4B, including Mac Studio M5 Ultra (256 GB), Mac Studio M5 Ultra (512 GB), MacBook Air 13" M3 (8 GB).
≤ 11 GB
≤ 11 GB tier →fits · 0.6 GB free · up to 32K ctx
68 GPUs · 1 Mac class · 3 device families
RTX 3080~155Simulate output on RTX 3080RTX 2080 Ti~126Simulate output on RTX 2080 TiRTX 3070 Ti~124Simulate output on RTX 3070 TiM3*8 GB~23Simulate output on iMac 24" M3 (8 GB)Orin Nano~21Simulate output on NVIDIA Jetson Orin Nano 8GB (Super)iPhone~15Simulate output on iPhone 1720 GB
fits · 16.6 GB free · up to 256K ctx
1 GPU
RX 7900 XT~151Simulate output on RX 7900 XT32–40 GB
fits · 28.6 GB free · up to 256K ctx
1 GPU · 3 Mac classes · 9 device families
RTX 5090~365Simulate output on RTX 5090M5 Max*36 GB~101Simulate output on Mac Studio M5 Max (36 GB)M632 GB~37Simulate output on iMac 24" M6 (32 GB)RTX Spark N1X~61Simulate output on NVIDIA RTX Spark N1X Laptop (32 GB)Orin~42Simulate output on NVIDIA Jetson AGX Orin 32GB
By VRAM tier · typical GPUs and Macs at Q4_K_M
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.5 4BAlibaba | 98.5 | 3.4 GB | ~201 | same size |
| Gemma 3 4B ITGoogle | 64.0 | 5 GB | ~231 | same size |
| Translategemma 4B ITGoogle | — | 5 GB | ~231 | same size |
| Molmo2 4BAllen AI | — | 4.7 GB | ~193 | same size |
| Qwen3 VL 4B InstructAlibaba | — | 4.5 GB | ~208 | same size |
| Qwen3.6 27B MTPLX Optimized SpeedYoussofal | — | 4 GB | ~185 | same size |
Specifications
- Publisher
- InternLM
- Parameters
- 4.5B
- Architecture
- Qwen3_5ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-09-26
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Intern Decision 4B need?
Intern Decision 4B requires 3.2 GB of VRAM at Q4_K_M, or 9.6 GB at BF16. Full 262K context adds up to 5.2 GB (8.4 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 4.5B × 4.8 bits ÷ 8 = 2.7 GB
KV Cache + Overhead ≈ 0.5 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.
KV Cache + Overhead ≈ 5.7 GB (at full 262K context)
VRAM usage by quantization
What's the best quantization for Intern Decision 4B?
For Intern Decision 4B, Q4_K_M (3.2 GB) offers the best balance of quality and VRAM usage. Q5_K_M (3.7 GB) provides better quality if you have the VRAM. The smallest option is Q2_K at 2.4 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Intern Decision 4B on a Mac?
Yes — MacBook Air 13" M3 (8 GB) and 158 other Macs can run Intern Decision 4B. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 3.2 GB of usable unified memory (RAM minus macOS overhead).
Can I run Intern Decision 4B locally?
Yes — Intern Decision 4B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Intern Decision 4B?
At Q4_K_M, Intern Decision 4B can reach ~1505 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~205 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 ÷ 3.19 × 0.65 = ~1630 tok/s
Estimated speed at Q4_K_M (3.2 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Intern Decision 4B?
At Q4_K_M, the download is about 2.72 GB. The full-precision BF16 version is 9.08 GB. The smallest option (Q2_K) is 1.93 GB.
Which GPUs can run Intern Decision 4B?
114 consumer GPUs can run Intern Decision 4B at Q4_K_M. Top options include AMD Radeon RX 5600 XT, AMD Radeon RX 5700 XT, AMD Radeon RX 580 8GB, NVIDIA GeForce GTX 1050 Ti. 107 GPUs have plenty of headroom for comfortable inference.
Which devices can run Intern Decision 4B?
185 devices with unified memory can run Intern Decision 4B at Q4_K_M (3.2 GB), including AMD Ryzen 7 8845HS (Hawk Point) Laptop, AMD Ryzen AI 7 350 (Krackan Point) Laptop, AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10. Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.