impulse

Qwen2.5 Coder 32B Instruct

Open weights

Qwen2.5-Coder is the latest series of Code-Specific Qwen large language models (formerly known as CodeQwen).

API id Qwen2.5-Coder-32B-InstructReleased 11 Nov 2024 (Nov 2024)
Input / 1M
$0.06
at Nscale
Output / 1M
$0.20
at Nscale
Context window
33K
tokens
Max output
33K
tokens per reply
Hosts
5
5 listings
Released
Nov 2024
11 Nov 2024

Pricing

List prices at Nscale, in US dollars per million tokens.

TierInput / 1MOutput / 1M
Standard$0.06$0.20

What it costs a month

At Nscale list prices, 30 days.

  • Chat assistant5,000 requests a day, 2K in / 0.5K out$33.00
  • RAG over documents2,000 requests a day, 8K in / 0.6K out$36.00
  • Coding agent1,000 requests a day, 40K in / 2K out$84.00
  • Overnight batch20,000 requests a day, 4K in / 0.4K out$192
Use your own numbers

Where to run it

5 hosts serve Qwen2.5 Coder 32B Instruct. The comparison is on blended price, 3 parts input to 1 part output.

HostIn / out per 1M
NscaleInference
Qwen2.5-Coder-32B-Instruct
$0.06 / $0.20
HyperbolicInference
Qwen2.5-Coder-32B-Instruct
$0.12 / $0.30
+74% vs headline
qwen2.5-coder-32b-instruct
$0.66 / $1.00
+684% vs headline
Qwen2.5-Coder-32B-Instruct
$0.87 / $0.87
+816% vs headline
Fireworks AIInference
qwen2p5-coder-32b-instruct
$0.90 / $0.90
+847% 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: Apache-2.0
Runs in about 78.6 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 Nscale.

Smaller Alibaba Qwen models

No cheaper Alibaba Qwen model keeps its key capabilities.

Other frontier labs

No cheaper match from another frontier lab.

Open-weight picks

Benchmarks

Scores published on the model card.

  • Aider Polyglot8.0%
  • BBH52.3%
  • GPQA13.2%
  • IFEval72.7%
  • MATH Lvl 549.5%
  • MMLU-Pro37.9%
  • MUSR13.7%
  • Open LLM Average39.9%

Model line

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

  1. Nov 2024Qwen25 Coder 32B InstructLatest$0.05 / $0.10
  2. Nov 2024Qwen2.5 Coder 32B InstructThis page$0.06 / $0.20

Call it

Using Nscale's API.

Python
from litellm import completion

response = completion(
    model="nscale/Qwen2.5-Coder-32B-Instruct",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Quick answers

How much does Qwen2.5 Coder 32B Instruct cost?

On Nscale, Qwen2.5 Coder 32B Instruct costs $0.06 per million input tokens and $0.20 per million output tokens.

Which APIs offer Qwen2.5 Coder 32B Instruct?

Qwen2.5 Coder 32B Instruct runs on 5 APIs: Nscale, Hyperbolic, Cloudflare Workers AI, OVHcloud AI Endpoints and Fireworks AI.

What is Qwen2.5 Coder 32B Instruct's context window?

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

Is Qwen2.5 Coder 32B Instruct being retired?

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

Is there a newer version of Qwen2.5 Coder 32B Instruct?

Yes. Qwen25 Coder 32B Instruct is the newest model in this line, released on 11 Nov 2024.

What are cheaper alternatives to Qwen2.5 Coder 32B Instruct?

Models that cost less and keep most of what it can do: Ling 3.0 Flash on DeepInfra (5% less). Savings are on a mix of 3 input tokens to 1 output token.

When was Qwen2.5 Coder 32B Instruct released?

Alibaba Qwen released it on 11 Nov 2024.

What is Qwen2.5 Coder 32B Instruct's API model id?

Use Qwen2.5-Coder-32B-Instruct with Nscale. Other hosts use their own ids: qwen2.5-coder-32b-instruct on Cloudflare Workers AI, qwen2p5-coder-32b-instruct on Fireworks AI.

Sources

Raw catalog entry for Qwen2.5-Coder-32B-Instruct on Nscale
{
  "input_cost_per_token": 6e-8,
  "litellm_provider": "nscale",
  "mode": "chat",
  "output_cost_per_token": 2e-7,
  "source": "https://docs.nscale.com/docs/inference/serverless-models/current#chat-models"
}