ICoder 27B — Hardware Requirements & GPU Compatibility
ChatCode
ICoder 27B is a 27.4B-parameter open language model from i-Coder. It supports a context window of up to 262,144 tokens. At BF16 it needs about 55.46 GB of VRAM — see which GPUs and Macs can run it below.
ICoder 27B runs on a 64 GB+ Mac or a workstation GPU
Best local pick → Mac Pro M2 Ultra (64 GB) · BF16 · 55.9 GB · ~10 tok/s
Macs from 64 GB (BF16 · ~10 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run ICoder 27B
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 i-Coder/iCoder-27B --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 A100 80GB (24 GB headroom) · BF16
- Generation speed
- ~24 tok/s
- generation speed
- Cost per 1M output tokens
- $12.74
- per 1M output tokens
How Much VRAM Does ICoder 27B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 55.5 GB | 56.8 GB | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 55.9 GB |
Selected BF16
Brain floating point 1655.9GBat 16K context
- Weights
- 54.7 GB
- KV cache
- 0.9 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 31.9 GB more
Or .
Which Hardware Can Run ICoder 27B?
BF16 · 55.5 GBICoder 27B (BF16) requires 55.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 73+ GB is recommended. Using the full 262K context window can add up to 13.9 GB, bringing total usage to 69.3 GB. No consumer GPU has enough memory. 52 devices with unified memory can run ICoder 27B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Pro M2 Ultra (64 GB).
64–80 GB
fits · 8.1 GB free · up to 128K ctx
3 Mac classes · 1 device family
M2 Ultra*64 GB~10Simulate output on Mac Pro M2 Ultra (64 GB)M5 Max64 GB~8Simulate output on Mac Studio M5 Max (64 GB)Orin~2Simulate output on NVIDIA Jetson AGX Orin 64GB
By VRAM tier · typical GPUs and Macs at BF16
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Qwen3.8 27BAlibaba | 122.0 | 17.9 GB | ~38 | same size |
| Qwen3.6 27BAlibaba | 111.2 | 17.9 GB | ~38 | same size |
| Qwen3.5 27BAlibaba | 103.8 | 17.9 GB | ~38 | same size |
| Diffusiongemma 26B A4B ITGoogle | — | 16.5 GB | ~171 | same size |
| Gemma 4 26B A4BGoogle | — | 16.5 GB | ~171 | same size |
| Medgemma 27B ITGoogle | — | 21.7 GB | ~36 | same size |
Specifications
- Publisher
- i-Coder
- Parameters
- 27.4B
- Architecture
- Qwen3_5ForConditionalGeneration
- Context Length
- 262,144 tokens
- Vocabulary Size
- 248,320
- Release Date
- 2026-08-21
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does ICoder 27B need?
ICoder 27B requires 55.5 GB of VRAM at BF16. Full 262K context adds up to 13.9 GB (69.3 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 27.4B × 16 bits ÷ 8 = 54.7 GB
KV Cache + Overhead ≈ 0.8 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 ≈ 14.6 GB (at full 262K context)
VRAM usage by quantization
Can NVIDIA GeForce RTX 5090 run ICoder 27B?
No — ICoder 27B requires at least 55.5 GB at BF16, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
Can I run ICoder 27B on a Mac?
Yes — Mac Pro M2 Ultra (64 GB) and 41 other Macs can run ICoder 27B. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 55.5 GB of usable unified memory (RAM minus macOS overhead).
Can I run ICoder 27B locally?
Yes — ICoder 27B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is ICoder 27B?
At BF16, ICoder 27B can reach ~87 tok/s on AMD Instinct MI350X. 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 ÷ 55.5 × 0.65 = ~94 tok/s
Estimated speed at BF16 (55.5 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of ICoder 27B?
At BF16, the download is about 54.71 GB.
Which GPUs can run ICoder 27B?
No single consumer GPU has enough VRAM to run ICoder 27B at BF16 (55.5 GB). Multi-GPU or professional hardware is required.
Which devices can run ICoder 27B?
52 devices with unified memory can run ICoder 27B at BF16 (55.5 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (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.