[whatmodel.sh](../../README.md) / DeepSeek-V2 Lite Chat (16B-A2.4B)
# DeepSeek-V2 Lite Chat (16B-A2.4B)
DeepSeek · released 2024-05 · 15.71B params (2.4B active) · deepseek
**score 39.8** ±1.1 · #571 of 732 · from 15 benchmark results across 5 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks DeepSeek-V2 Lite Chat (16B-A2.4B) among every model on that task or test, worst to best.
task vs the others rank counts tests
coding ▁▁▁▃●▇█▇▅▄▃▁ #390 of 469 29% 7
math █··········● #1 of 2* 13% 2
coding · #390 of 469 · 29% of its score
Beats [LFM2.5 8B-A1B](lfm2-5-8b-a1b.md), [DeepSeek R1 Distill Qwen 7B](deepseek-r1-distill-qwen-7b.md), [Qwen3.5 2B](qwen3-5-2b.md) here, though they rank higher overall. Trails [Qwen2.5 1.5B Instruct](qwen2-5-1-5b.md), [WizardLM 13B V1.2](wizardlm-13b-v1-2.md), [Falcon3 3B Instruct](falcon3-3b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LiveCodeBench official (Jan 1–Nov… 3.6% ●▁·▁▂▄▆█▅▂·▂ beats 0/31 hard · avg 23% 10% sharp
HumanEval (maker) 57.3% ▁▁▁▁▂▂▆●█▅▂▁ beats 111/205 easy · avg 81% 7.8% sharp · at its level · small run
MBPP (maker) 45.8% ▁▁▁▄▅●▆▅▄▁▂▁ beats 27/86 easy · avg 68% 4.0% at its level · small run
CodeReviewQA: solution identifica… 20.5% ▃··▁▂▂●▂█▃▄▂ beats 15/47 medium · avg 51% 2.5% sharp · noisy
CodeReviewQA: automated code refi… 15.0% ▁▁▁▁▁▁●▂▃█▅▄ beats 10/47 hard · avg 32% 2.2% typical
CodeReviewQA: change localisation… 6.9% █▂▁▁▂●▂▃▅▅▅▇ beats 16/47 hard · avg 33% 1.8% sharp · noisy
CodeReviewQA: change type recogni… 76.8% ▁▁▁······▁▃● beats 38/47 easy · avg 67% 0.1% blunt · at its level · noisy
math · #1 of 2 · 13% of its score
test result vs the others in its run difficulty counts why
GSM8K (maker) 72.0% ▂▂▂▂▃▄▅▇●█▆▄ beats 121/206 very easy · avg 89% 8.4% sharp · at its level · small run
MATH (maker) 27.9% ▃▅▂▆▅▃●▅█▃▃▂ beats 16/30 medium · avg 45% 4.7% small run
other tests (6)
test result vs the others in its run difficulty counts why
LiveBench 2024-07-26 17.7% ▁▁▁▂●▃▂█▆▃▅▃ beats 11/61 medium · avg 38% 34% typical
BIG-Bench Hard (maker) 48.1% ▁▁▁▁▂▄●▆█▃▃▁ beats 55/155 medium · avg 71% 7.5% sharp · at its level · small run
MMLU (maker) 55.7% ▁▁▁▁▁▂▂▅●█▆▄ beats 71/214 easy · avg 76% 6.1% at its level · small run
C-Eval (maker) 60.1% ▁···▁▁▂▁●▂▃█ beats 8/33 medium · avg 72% 3.7% at its level · small run
CMMLU (maker) 62.5% ▅·●····█▃▅█▅ beats 3/14 easy · avg 74% 3.2% at its level · small run
Open PT-LLM Leaderboard average (… 63.3% ▃▁▁▁▁▁▁▁●█▅▁ beats 102/276 medium · avg 72% 2.8% at its level · noisy
source results share of the score
DeepSeek vendor-reported benchmark scor… 8 45% ▇▇▇▇▇
LiveBench model judgments 1 34% ▇▇▇▇
LiveCodeBench official website results 1 10% ▇
CodeReviewQA: The Code Review Comprehen… 4 7% ▇
Open Portuguese LLM Leaderboard 1 3% ▇
prior: pull toward the middle 1% ▇
weighted average = 39.8
Licences: LiveBench model judgments ([Apache-2.0 (stated in the LiveBench datasheet)](https://www.apache.org/licenses/LICENSE-2.0)) · LiveCodeBench official website results ([CC-BY-SA-4.0](../../licenses/lcb-sharealike-NOTICE.txt)) · CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run DeepSeek-V2 Lite Chat (16B-A2.4B) (Q4_K_M, needs ~10.4 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/DeepSeek-V2-Lite-Chat-GGUF/tree/9ee9f1f8f39f673e1022e176d586aef4228f630d))
```sh
llama cli -hf mradermacher/DeepSeek-V2-Lite-Chat-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf mradermacher/DeepSeek-V2-Lite-Chat-GGUF:Q4_K_M
```
Other versions: Q8_0 (~16.7 GB) `llama cli -hf mradermacher/DeepSeek-V2-Lite-Chat-GGUF:Q8_0`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run DeepSeek-V2 Lite Chat (16B-A2.4B) (Q4_K_M, needs ~10.4 GB, from [huggingface.co/mradermacher](https://huggingface.co/mradermacher/DeepSeek-V2-Lite-Chat-GGUF/tree/9ee9f1f8f39f673e1022e176d586aef4228f630d))
```sh
lms get https://huggingface.co/mradermacher/DeepSeek-V2-Lite-Chat-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q8_0 (~16.7 GB) `lms get https://huggingface.co/mradermacher/DeepSeek-V2-Lite-Chat-GGUF@Q8_0 -y` · 4bit (~8.8 GB, Mac only) `lms get https://huggingface.co/mlx-community/DeepSeek-V2-Lite-Chat-4bit-mlx -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run DeepSeek-V2 Lite Chat (16B-A2.4B) (Q4_K_M, needs ~9.93 GB, from [ollama.com](https://ollama.com/library/deepseek-v2/tags))
```sh
ollama run deepseek-v2:16b-lite-chat-q4_K_M
```
Other versions: Q8_0 (~19.35 GB) `ollama run deepseek-v2:16b-lite-chat-q8_0`
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run DeepSeek-V2 Lite Chat (16B-A2.4B) (4bit, needs ~8.8 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/DeepSeek-V2-Lite-Chat-4bit-mlx/tree/a33a8ad64477b9ca7a65b6dc1d8c06c997f87216))
```sh
mlx_lm.chat --model mlx-community/DeepSeek-V2-Lite-Chat-4bit-mlx
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/DeepSeek-V2-Lite-Chat-4bit-mlx
```
Model repository and documentation: [huggingface.co/deepseek-ai](https://huggingface.co/deepseek-ai/DeepSeek-V2-Lite-Chat)
[interactive version](https://whatmodel.sh/model/deepseek-v2-lite-chat) · [JSON](https://whatmodel.sh/api?format=json&all=1)