impulse

DeepSeek V2.5

Open weights
API id deepseek-v2p5
Input / 1M
$1.20
at Fireworks AI
Output / 1M
$1.20
at Fireworks AI
Context window
33K
tokens
Max output
33K
tokens per reply
Hosts
1
1 listings
Released
-
date not published

Pricing

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

TierInput / 1MOutput / 1M
Standard$1.20$1.20

What it costs a month

At Fireworks AI list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$450
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$619
  • Coding agent1,000 requests a day, 40K in / 2K out$1,512
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$3,168
Use your own numbers

Where to run it

1 host serve DeepSeek V2.5. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
Fireworks AIInference
deepseek-v2p5
$1.20 / $1.20

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

No host sells it for less than Fireworks AI.

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
    29% less
  • GPT-5.4 Nano
    OpenAI · on OpenAI
    $0.20 / $1.25
    Has every key capability
    61% less
  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    83% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    83% less

Open-weight picks

  • Motif 3
    Independent · on ai&
    $0.50 / $2.00
    Has every key capability
    27% less
  • GLM 5.3
    Zhipu AI · on CLF AI Gateway
    $0.56 / $1.76
    Has every key capability
    28% less
  • Nemotron 3 Ultra 550B A55B
    NVIDIA · on Perplexity
    $0.25 / $2.50
    Has every key capability
    32% less
  • DeepSeek V4 Pro
    DeepSeek · on CLF AI Gateway
    $0.528 / $1.58
    Has every key capability
    34% 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 V3$0.89 / $0.89
  5. UndatedDeepSeek V2.5This page$1.20 / $1.20

Call it

Using Fireworks AI's API.

Python
from litellm import completion

response = completion(
    model="fireworks_ai/deepseek-v2p5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does DeepSeek V2.5 cost?

On Fireworks AI, DeepSeek V2.5 costs $1.20 per million input tokens and $1.20 per million output tokens.

Which APIs offer DeepSeek V2.5?

DeepSeek V2.5 runs on 1 API: Fireworks AI.

What is DeepSeek V2.5's context window?

33K tokens, with up to 33K tokens in one reply.

Is DeepSeek V2.5 being retired?

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

Is there a newer version of DeepSeek V2.5?

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

What are cheaper alternatives to DeepSeek V2.5?

Models that cost less and keep most of what it can do: Gemini 3.5 Flash Lite on Google AI Studio (29% less), GPT-5.4 Nano on OpenAI (61% less) and Motif 3 on ai& (27% less). Savings are on a mix of 3 input tokens to 1 output token.

What is DeepSeek V2.5's API model id?

Use deepseek-v2p5 with Fireworks AI.

Sources

Raw catalog entry for deepseek-v2p5 on Fireworks AI
{
  "max_input_tokens": 32768,
  "max_output_tokens": 32768,
  "input_cost_per_token": 0.0000012,
  "output_cost_per_token": 0.0000012,
  "litellm_provider": "fireworks_ai",
  "mode": "chat"
}