impulse

DeepSeek V3

Open weights
API id deepseek_v3
Input / 1M
$0.89
at Novita AI
Output / 1M
$0.89
at Novita AI
Context window
64K
tokens
Max output
16K
tokens per reply
Hosts
10
11 listings
Released
-
date not published

Pricing

List prices at Novita AI, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.89$0.89

What it costs a month

At Novita AI list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$334
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$459
  • Coding agent1,000 requests a day, 40K in / 2K out$1,121
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$2,350
Use your own numbers

Where to run it

10 hosts serve DeepSeek V3. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
Lambda AIInference
0324
deepseek-v3-0324
$0.20 / $0.60
−66% vs headline
HyperbolicInference
0324
DeepSeek-V3-0324
$0.40 / $0.40
−55% vs headline
DeepInfraInference
0324
DeepSeek-V3-0324
$0.24 / $0.90
−54% vs headline
GMI CloudInference
0324
DeepSeek-V3-0324
$0.28 / $0.88
−52% vs headline
0324
DeepSeek-V3-0324
$0.50 / $1.50
−16% vs headline
BasetenInference
0324
DeepSeek-V3-0324
$0.77 / $0.77
−13% vs headline
Novita AIInference
deepseek_v3
$0.89 / $0.89
Fireworks AIInference
0324
deepseek-v3-0324
$0.90 / $0.90
+1% vs headline
CrusoeInference
0324
DeepSeek-V3-0324
$1.50 / $1.50
+69% vs headline
0324
deepseek-v3-0324
$1.14 / $4.56
+124% vs headline
Show 1 more listings: regions, snapshots and variants
Novita AI
0324
deepseek-v3-0324
$0.27 / $1.12
−46% vs headline

Capabilities

Core

  • Streaming
  • Function calling
  • JSON mode
  • Structured outputs

Multimodal

  • Vision
  • Image input
  • Audio
  • Image generation
  • Video

Advanced

  • Reasoning
  • Tool calling
  • MCP
  • Prompt caching
  • Batch API
  • Embeddings
Takes in
Text
Produces
Text
Open weights
License: DeepSeek License (MIT code)
Runtimes: SGLang, TGI, vLLM

Cheaper alternatives

Current models that cost less and keep most of what this one can do: tools, vision, reasoning, structured output, caching and enough context.

Same model, cheaper host

  • Lambda AI
    deepseek-v3-0324
    $0.20 / $0.60
    66% less
  • Hyperbolic
    DeepSeek-V3-0324
    $0.40 / $0.40
    55% less
  • DeepInfra
    DeepSeek-V3-0324
    $0.24 / $0.90
    54% less
  • GMI Cloud
    DeepSeek-V3-0324
    $0.28 / $0.88
    52% less
  • Novita AI
    deepseek-v3-0324
    $0.27 / $1.12
    46% less

Smaller DeepSeek models

No cheaper DeepSeek model keeps its key capabilities.

Other frontier labs

  • Gemini 3.5 Flash Lite
    Google · on Google AI Studio
    $0.30 / $2.50
    Has every key capability
    4% less
  • GPT-5.4 Nano
    OpenAI · on OpenAI
    $0.20 / $1.25
    Has every key capability
    48% less
  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    78% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    78% less

Open-weight picks

  • Motif 3
    Independent · on ai&
    $0.50 / $2.00
    Has every key capability
    2% less
  • GLM 5.3
    Zhipu AI · on CLF AI Gateway
    $0.56 / $1.76
    Has every key capability
    3% less
  • Nemotron 3 Ultra 550B A55B
    NVIDIA · on Perplexity
    $0.25 / $2.50
    Has every key capability
    9% less
  • DeepSeek V4 Pro
    DeepSeek · on CLF AI Gateway
    $0.528 / $1.58
    Has every key capability
    11% less

Model line

Newer and older versions, at each lab's own price.

  1. UndatedDeepSeek V3.2.251201LatestNot listed
  2. Dec 2025DeepSeek V3.2$0.28 / $0.40
  3. UndatedDeepSeek V3.1$0.25 / $0.95
  4. UndatedDeepSeek V3This page$0.89 / $0.89
  5. UndatedDeepSeek V2.5$1.20 / $1.20

Call it

Using Novita AI's API.

Python
from litellm import completion

response = completion(
    model="novita/deepseek_v3",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does DeepSeek V3 cost?

On Novita AI, DeepSeek V3 costs $0.89 per million input tokens and $0.89 per million output tokens. The cheapest host is Lambda AI (0324), at $0.20 per million input tokens and $0.60 per million output tokens.

Which APIs offer DeepSeek V3?

DeepSeek V3 runs on 10 APIs: Lambda AI, Hyperbolic, DeepInfra, GMI Cloud, Nebius AI Studio, Baseten, Novita AI, Fireworks AI, Crusoe and Azure AI Foundry.

What is DeepSeek V3's context window?

64K tokens, with up to 16K tokens in one reply.

Is DeepSeek V3 being retired?

No retirement date is scheduled for it on any API we track.

Is there a newer version of DeepSeek V3?

Yes. DeepSeek V3.2.251201 is the newest model in this line.

What are cheaper alternatives to DeepSeek V3?

Models that cost less and keep most of what it can do: Gemini 3.5 Flash Lite on Google AI Studio (4% less), GPT-5.4 Nano on OpenAI (48% less) and Motif 3 on ai& (2% less). DeepSeek V3 itself costs 66% less on Lambda AI (0324). Savings are on a mix of 3 input tokens to 1 output token.

What is DeepSeek V3's API model id?

Use deepseek_v3 with Novita AI. Other hosts use their own ids: deepseek-v3-0324 on Lambda AI, DeepSeek-V3-0324 on Hyperbolic, DeepSeek-V3-0324 on DeepInfra, DeepSeek-V3-0324 on GMI Cloud.

Sources

Raw catalog entry for deepseek_v3 on Novita AI
{
  "input_cost_per_token": 8.900000000000001e-7,
  "litellm_provider": "novita",
  "max_input_tokens": 64000,
  "max_output_tokens": 16000,
  "mode": "chat",
  "output_cost_per_token": 8.900000000000001e-7,
  "source": "https://api.novita.ai/v3/openai/models",
  "supports_function_calling": true,
  "supports_tool_choice": true,
  "supports_vision": false
}