impulse

Qwen 2.5 72B Instruct

Open weights

Qwen2.5 72B is the latest series of Qwen large language models.

API id qwen-2.5-72b-instructReleased 19 Sep 2024 (Sep 2024)
Input / 1M
$0.38
at Novita AI
Output / 1M
$0.40
at Novita AI
Context window
32K
tokens
Max output
8K
tokens per reply
Hosts
1
1 listings
Released
Sep 2024
19 Sep 2024

Pricing

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

TierInput / 1MOutput / 1M
Standard$0.38$0.40

What it costs a month

At Novita AI list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$144
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$197
  • Coding agent1,000 requests a day, 40K in / 2K out$480
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$1,008
Use your own numbers

Where to run it

1 host serve Qwen 2.5 72B Instruct. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
Novita AIInference
qwen-2.5-72b-instruct
$0.38 / $0.40

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
Features
Parallel tool calls
Open weights
License: Apache-2.0
Runs in about 174.5 GB of GPU memory
Runtimes: Ollama, SGLang, llama.cpp, 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 Novita AI.

Smaller Alibaba Qwen models

No cheaper Alibaba Qwen model keeps its key capabilities.

Other frontier labs

  • Claude Haiku 5.5
    Anthropic · on Anthropic
    $0.10 / $0.50
    Has every key capability
    48% less
  • GPT-6 Luna
    OpenAI · on OpenAI
    $0.10 / $0.50
    Has every key capability
    48% less
  • GPT-4o Mini
    OpenAI · on OpenAI
    $0.15 / $0.60
    Has every key capability
    32% less

Open-weight picks

Benchmarks

Scores published on the model card.

  • BBH61.9%
  • GPQA16.7%
  • IFEval86.4%
  • MATH Lvl 559.8%
  • MMLU-Pro51.4%
  • MUSR11.7%
  • Open LLM Average48.0%

Model line

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

  1. Sep 2024Qwen2.5 72B InstructLatest$0.12 / $0.30
  2. Sep 2024Qwen 2.5 72B InstructThis page$0.38 / $0.40
  3. UndatedQwen2 72B Instruct$0.90 / $0.90

Call it

Using Novita AI's API.

Python
from litellm import completion

response = completion(
    model="novita/qwen-2.5-72b-instruct",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Qwen 2.5 72B Instruct cost?

On Novita AI, Qwen 2.5 72B Instruct costs $0.38 per million input tokens and $0.40 per million output tokens.

Which APIs offer Qwen 2.5 72B Instruct?

Qwen 2.5 72B Instruct runs on 1 API: Novita AI.

What is Qwen 2.5 72B Instruct's context window?

32K tokens, with up to 8K tokens in one reply.

Is Qwen 2.5 72B Instruct being retired?

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

Is there a newer version of Qwen 2.5 72B Instruct?

Yes. Qwen2.5 72B Instruct is the newest model in this line, released on 19 Sep 2024.

What are cheaper alternatives to Qwen 2.5 72B Instruct?

Models that cost less and keep most of what it can do: Claude Haiku 5.5 on Anthropic (48% less), GPT-6 Luna on OpenAI (48% less) and Qwen3.5 397B A17B on AIHubMix (4% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Qwen 2.5 72B Instruct released?

Alibaba Qwen released it on 19 Sep 2024.

What is Qwen 2.5 72B Instruct's API model id?

Use qwen-2.5-72b-instruct with Novita AI.

Sources

Raw catalog entry for qwen-2.5-72b-instruct on Novita AI
{
  "litellm_provider": "novita",
  "mode": "chat",
  "input_cost_per_token": 3.8e-7,
  "output_cost_per_token": 4e-7,
  "max_input_tokens": 32000,
  "max_output_tokens": 8192,
  "supports_function_calling": true,
  "supports_parallel_function_calling": true,
  "supports_tool_choice": true,
  "supports_system_messages": true,
  "supports_response_schema": true
}