DeepSeek V3

Featured Latest
DeepSeek DeepSeek V3 Generally Available Dec 2024

DeepSeek's flagship MoE model — 671B total / 37B active params, 128K context, and near-GPT-4 performance at a fraction of the cost. MIT licence with 4K+ HuggingFace likes.

Context
128K
tokens
Input
$0.32
per MTok
Output
$0.89
per MTok

About

DeepSeek-V3 — AI tool

Modalities

Input
Text Code
Output
Text Code

Code Examples

curl https://openrouter.ai/api/v1/chat/completions \
  -H "Authorization: Bearer $OPENROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-ai/DeepSeek-V3",
    "messages": [
      { "role": "user", "content": "Explain quantum entanglement in one sentence." }
    ]
  }'

API Parameters

Name Type Description
frequency_penalty number Penalize tokens by their frequency so far. Positive values reduce repetition.
logit_bias object Map of token-id to bias (-100…100) added to the logit before sampling.
max_tokens deprecated integer Deprecated. Use max_completion_tokens.
min_p unknown
presence_penalty number Penalize tokens that have appeared at all so far. Positive values encourage new topics.
repetition_penalty number Penalize repeated tokens (>1.0 reduces repetition, <1.0 encourages it).
response_format one of Constrain output to a JSON schema or an enum (structured outputs).
seed integer Deterministic seed for sampling. Same seed + same prompt produces identical output.
stop array Up to 4 sequences where the API will stop generating tokens.
structured_outputs boolean Enable JSON-schema-constrained output.
temperature number Sampling temperature; higher values produce more random output. 0 is deterministic.
tool_choice one of Controls which (if any) tool is called: "none", "auto", "required", or a specific tool.
tools array List of tools (functions) the model may call.
top_k integer Limit sampling to the top-k most likely tokens at each step.
top_p number Nucleus sampling: consider only tokens whose cumulative probability ≥ top_p.

Standard OpenAI-compatible parameters. Consult the provider docs for model-specific behaviour.

Benchmark Scores

Benchmark Score
MMLU 88.5%
MMLU-Pro 75.9%
GPQA Diamond 59.1%
HumanEval 65.2%

Performance

60
tok / sec
output speed

Source: DeepSeek V3 technical report + Artificial Analysis, April 2026

Strengths & Limitations

Best For
Near-GPT-4 MMLU at near-open-source pricing
MIT licence
MoE architecture reduces inference cost
4K HuggingFace likes
730K downloads
Limitations
671B total params require specialised MoE infrastructure
No vision modality
Inferior to R1 on reasoning tasks

Tags

Open WeightsMoECodingReasoningCost Efficient