DeepSeek-Coder-V2-Lite-Instruct

Deepseek-Coder family · 16.0B params

1recorded run

237.3best tok/s

1verified

Evidence: 1 sourced run · 1 independent source · first recorded Aug 11, 2026. What these labels mean.

What the DeepSeek-Coder-V2-Lite-Instruct numbers mean

The fastest verified result on our board is 237.3 output tok/s on an RTX PRO 6000 Blackwell at FP8 with a 32768-token context. That is about 4.2 ms per generated token — 500 tokens takes roughly 2.1 seconds.

How much VRAM does DeepSeek-Coder-V2-Lite-Instruct need?

DeepSeek-Coder-V2-Lite-Instruct has 16.0B parameters. These estimates use the same formula as the VRAM calculator: weights, plus about 15% for a moderate-context KV cache and runtime overhead, plus a small engine floor.

Quantisation	Estimated VRAM	Smallest card that fits
Q4_K_M (4-bit)	11.2 GB	16 GB
Q8_0 (8-bit)	20.3 GB	24 GB
FP16 (16-bit)	37.6 GB	48 GB

Long contexts and large batches raise the real number — read the smallest card as "possible with a reduced context". Measured throughput is on the benchmark board.

Fastest GPUs for this model (best tok/s)

RTX PRO 6000 Blackwell237.3

Models with a similar footprint

Model	Params	Q4_K_M weights
DeepSeek-R1-Distill-Qwen-14B	15B	8.4 GB
Nemotron-Cascade-14B-Thinking	15B	8.4 GB
phi-4	15B	8.4 GB
Phi-4-reasoning	15B	8.4 GB
Phi-4-reasoning-plus	15B	8.4 GB
Qwen2.5-Coder-14B-Instruct	15B	8.4 GB

Ranked by distance from DeepSeek-Coder-V2-Lite-Instruct's parameter count. Full benchmark board.

All benchmarks Methodology

GPU	Model	Quant	Backend	Ctx	tok/s	prompt tok/s	VRAM GB	Source
RTX PRO 6000 Blackwell · 96 GB	DeepSeek-Coder-V2-Lite-Instruct	FP8	vllm 0.26.0	32768	237.3 ✓	—	—	src