DeepSeek V3

The open-weight 671B-param MoE that put DeepSeek on the frontier map.

Context window

128K tokens

Input / 1M tokens

Free

Output / 1M tokens

Free

Provider

DeepSeek

Open-weight (MIT license) - free to self-host. Hosted API access is also available from third parties (e.g. OpenRouter lists ~$0.20/1M input and ~$0.80/1M output). Note: on the official DeepSeek platform the 'deepseek-chat' model id that mapped to the V3 line is being deprecated on 2026-07-24 in favor of the V4 models. · Data verified 2026-07-02

DeepSeek V3 is an open-weight mixture-of-experts model (671B total / 37B active) first released December 26, 2024 under the MIT license, pre-trained on nearly 15 trillion tokens. It offers a 128K-token context window and strong instruction-following and coding performance, competing with both open and closed models of its era. It has since been superseded on DeepSeek's own platform by the V4 family, but the open weights remain widely used and hosted across providers.

Capability index

Relative estimates (0-100) to place this model against its peers, grounded in published benchmarks.

Coding

Reasoning

Math

Multimodal

Long context

Speed

Cost efficiency

How to access it

Download open weights from Hugging Face (deepseek-ai/DeepSeek-V3) to self-host, or access via third-party inference providers such as OpenRouter, Together, and DeepInfra. Ollama/vLLM support local serving.

Get access →Documentation →

Strengths

✓Open weights under MIT license - fully self-hostable and fine-tunable
✓Strong general-purpose reasoning, coding, and instruction following for its generation
✓128K context window suitable for long documents
✓Very low cost via third-party API providers
✓Large, well-supported ecosystem (many hosts, tooling, fine-tunes)

Best for developers who...

Cost-efficient open-weight general useSelf-hosting and fine-tuningCoding and instruction-followingReproducible pinned deployments

When to choose it (and when not to)

Reach for DeepSeek V3 when...

→You want a proven, stable open model with broad ecosystem support
→You need to self-host or fine-tune without licensing friction
→Cost is critical and you don't need V4's 1M context or top scores
→You want reproducible open-weight behavior pinned to a known version

Look elsewhere if...

✕You need the latest frontier performance or a 1M context (use DeepSeek V4)
✕You rely on the official DeepSeek 'deepseek-chat' alias long-term (being deprecated 2026-07-24)
✕You need multimodal input
✕You need very long outputs beyond typical limits

How to use it

›Provide clear, explicit instructions - V3 follows structured prompts well
›Use system prompts to set role and output format (e.g. JSON)
›For coding, include full context and desired language/framework explicitly
›Pin to a specific hosted V3 build (e.g. the 0324 revision) for reproducibility

Quickstart

Python

from openai import OpenAI

# Example via OpenRouter (third-party host of the open weights)
client = OpenAI(
    api_key="YOUR_OPENROUTER_API_KEY",
    base_url="https://openrouter.ai/api/v1",
)

resp = client.chat.completions.create(
    model="deepseek/deepseek-chat",
    messages=[{"role": "user", "content": "Write a Python function to parse an ISO date."}],
)
print(resp.choices[0].message.content)

The open weights (deepseek-ai/DeepSeek-V3) can also be run locally with Ollama or vLLM instead of a hosted API.

API model id: deepseek-chat