How much VRAM does StarCoder2 15B need?

About 10.2 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 15.0 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.

Will StarCoder2 15B run on a 12GB GPU?

Yes — at Q4_K_M and 4K context it needs about 10.2 GB, which leaves 1.8 GB spare on a RTX 3060 12G, or on any of the other 12 GB cards here (10 in all). That is the smallest size in this index that clears it comfortably; 65 of 78 do.

Which quantization of StarCoder2 15B should I use?

Q4_K_M has the lowest published quality loss (3.2%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, GPTQ INT4.

StarCoder2 15B

15B

BigCode

⚠ Superseded · Prefer Qwen2.5-Coder 7B Instruct

This index would suggest Qwen2.5-Coder 7B Instruct instead today — longer context (16K → 128K tokens). The numbers below are still accurate; they are just for a model you probably should not start with. Compare the two →

BigCode's open code model trained on 600+ languages. Great for polyglot dev.

Consumer GPU

16K

Max Context

Quant Variants

GGUF Q4_K_M

Best Quality

96.8%

Accuracy Retained

Calculate VRAM Hugging Face Compare

Quantization Variants

Per-quant VRAM, quality loss, and inference speed on RTX 4090

Measured = run on this site · Estimated = not run here: calculated, or a figure this site has not verified · Community = public reports

Format	Level	BPW	VRAM	PPL Loss	Speed	Source	Actions
GGUF	Q4_K_M	4.85	10.5 GB	3.2%	92 tok/s	Estimated	Calc HF
GPTQ	INT4	4	9.2 GB	4.8%	115 tok/s	Estimated	Calc HF

Similar models

Compare with Qwen2.5 14B

14B

Qwen2.5 14B Instruct

Alibaba Qwen2.5

Consumer GPUMac / Apple Silicon

9.2 GBmin VRAM98.6%accuracy

The sweet spot between performance and resource usage. 16GB VRAM with Q4.

16B-A2.4B

DeepSeek-Coder-V2-Lite Instruct

DeepSeek

Consumer GPUMac / Apple Silicon

9.8 GBmin VRAM99.8%accuracy

MoE architecture coding model. Active params ~2.4B, total ~16B. Exceptional code quality. Its attention is MLA: a GGUF converted by current llama.cpp (tensors attn_k_b / attn_v_b) caches about 1 GB at 32K, which is what the estimates here assume. An older conversion with a single attn_kv_b tensor caches full keys and values instead — about 8.4 GB at 32K.

14B

DeepSeek-R1-Distill-Qwen-14B

DeepSeek

Consumer GPU

9.2 GBmin VRAM98.0%accuracy

R1 reasoning distilled into 14B. Huge community interest; excellent chain-of-thought.

14B

Phi-3 Medium 14B Instruct

Microsoft Phi

Consumer GPUMac / Apple Silicon

8.8 GBmin VRAM99.2%accuracy

Microsoft's mid-size Phi-3. Excellent quality-per-GB on 16GB cards.

Running StarCoder2 15B locally

At Q4_K_M and 4K of context, StarCoder2 15B needs about 10.2 GB — 8.6 GB of weights, 0.63 GB of KV cache and a 0.9 GB activation buffer. The smallest cards in this index that clear that comfortably are 12 GB ones — 10 of them, from the RTX 3060 12G up — and 65 of the 78 cards here do. These are calculated figures, not measurements: the estimate stops counting a card as comfortable at 88% of its VRAM, which is roughly the room a desktop session needs.

What longer context costs

Going from 4K to 32K adds about 4.37 GB, taking the total to 15.0 GB. Weights do not move with context — only the KV cache does, and it grows linearly, so this is the number to watch when planning for long documents. The model's native window is 16K; holding all of it at Q4_K_M would need about 12 GB.

Which build to download

This index tracks 2 formats for it — GGUF, GPTQ — across 2 levels. Q4_K_M carries the lowest published perplexity loss at 3.2%. The fastest level listed is GPTQ INT4 at about 115 tok/s on an RTX 4090, batch 1 — an estimate, not a run on this site's hardware. GGUF runs on llama.cpp and Ollama across NVIDIA, AMD and Apple silicon; AWQ and GPTQ target vLLM on CUDA and ROCm; EXL2 is ExLlamaV2 — CUDA only, and now archived, so local use rather than serving.

Common questions

How much VRAM does StarCoder2 15B need?: About 10.2 GB at Q4_K_M with 4K of context and batch 1, rising to roughly 15.0 GB at 32K. That figure is weights plus KV cache plus a 10% activation buffer, calculated from the model's architecture rather than measured on a card.
Will StarCoder2 15B run on a 12GB GPU?: Yes — at Q4_K_M and 4K context it needs about 10.2 GB, which leaves 1.8 GB spare on a RTX 3060 12G, or on any of the other 12 GB cards here (10 in all). That is the smallest size in this index that clears it comfortably; 65 of 78 do.
Which quantization of StarCoder2 15B should I use?: Q4_K_M has the lowest published quality loss (3.2%), and Q4_K_M is the level most people run. All 2 levels in the index are Q4_K_M, GPTQ INT4.

Where this model fits

Sized at Q4_K_M with a 4K context window, smallest card first. Comfortable means the estimate uses at most 88% of the memory.

Runs comfortably on: RTX 3060 12G12GB · +1.8 Radeon RX 6700 XT12GB · +1.8 Radeon RX 7700 XT12GB · +1.8 Arc B580 12G12GB · +1.8 RTX 4070 Ti12GB · +1.8
Tight but possible: Arc B570 10G10GB RTX 3080 10G10GB
Just misses: RTX 4060+2.2GB over RTX 4060 Ti 8G+2.2GB over
Ships in: GGUF GPTQ
Compared here: GGUF vs GPTQ
Try it yourself: Size it in the calculator Generate the run command

How to actually run this

Deployment guides for this model and this class of hardware.

This page's figures change when the model or the runtime does.Last updated 2026-10-20 RSS → /feed.xml