StarCoder15B params16k context

StarCoder 2 15B locally

StarCoder 2 15B is an open-weight LLM from the StarCoder family with 15B parameters. Main use: code and developer agents. Detected minimum hardware: RTX 3080 10GB (10 GB).

Technical facts
Parameters15B
Max context16k
Q4_K_M9.4 GB
Q5_K_M11.5 GB
Q816.8 GB
FP1633.5 GB
FamilyStarCoder
Last sync2026-05-12

Available quantizations

Q4_K_M
9.4GB

Acceptable. Good compromise when VRAM is limited.

Q5_K_M
11.5GB

Good quality. Sweet spot for size and precision.

Q8
16.8GB

Near-FP16 quality. Comfortable for production.

FP16
33.5GB

Reference precision. Maximum quality, doubled VRAM.

Compatible GPUs

GPUs that can run StarCoder 2 15B on a single card, ranked by VRAM headroom.

Recommended multi-GPU rigs

For StarCoder 2 15B at higher quantization or with more context, a multi-GPU rig gives more headroom.

Recommended rig

2× GTX 1060 6GB

StarCoder 2 15B with Ubuntu, vLLM, Open WebUI and the model already downloaded.

Configure

Similar models

VRAM estimates: parameters x bits/8 plus margin. Real performance varies by engine, context length and batch size.
sync: 2026-05-12