Alibaba (Qwen) · language model
Qwen2.5 VL 72B Instruct: pricing, context and benchmarks
Qwen2.5 VL 72B Instruct is Alibaba (Qwen)'s model released on February 1, 2025. It costs $0.8 per million input tokens and $1 per million output tokens, with a 128K-token context window. No independent score has been published for it so far.
What you need to know about Qwen2.5 VL 72B Instruct
Alibaba (Qwen), based in China, put Qwen2.5 VL 72B Instruct online on February 1, 2025. It accepts text and images as input and produces text. Its context window reaches 128,000 tokens, larger than 19 % of the models in the registry, and it can write up to 115,200 tokens in one go.
Released more than 1 year ago, it belongs to a generation most vendors have since superseded.
Qwen2.5 VL 72B Instruct pricing
On OpenRouter, Qwen2.5 VL 72B Instruct costs $0.8 per million input tokens and $1 per million output tokens. Counting three tokens read for every token written, its average price is $0.85 per million, more expensive than 53 % of paid models (1.1 times the median of $0.76).
A concrete example: a task that sends 2,000 tokens (about 1,500 words) and receives a 500-token answer costs $0.0021. At 1,000 tasks a day, budget $63 a month. The price calculator redoes this with your own volumes.
Tokens already sent and cached are billed at $0.4 per million.
Performance and benchmarks
Epoch AI does not yet publish a capabilities index for Qwen2.5 VL 72B Instruct. We therefore show no made-up performance rating: only price, context and features count towards its score.
Scores published by Epoch AI on the benchmarks the site tracks:
| Benchmark | What it measures | Score | Date |
|---|---|---|---|
| OSWorld | operating a desktop computer | 5 % | Sep 19, 2024 |
Features, privacy and local use
- It does not support tool calling: it is not suited to driving an AI agent that has to act.
- It can answer in an enforced JSON format (structured outputs).
- Its knowledge stops at June 2024.
Its weights are open and downloadable from Hugging Face: you can run it on your own machines without sending your data to a third party.
Other Alibaba (Qwen) models
| Model | Released | Avg. price | ECI |
|---|---|---|---|
| Qwen3.8 Max Prime | Sep 23, 2026 | $6 | — |
| Qwen3.8 Omni Flash | Sep 21, 2026 | $0.23 | — |
| Qwen3.8 Max (0902) | Sep 3, 2026 | $3 | 155.1 |
| Qwen3.8 Flash | Aug 26, 2026 | $0.23 | — |
| Qwen3.8 27B | Aug 14, 2026 | $0.78 | 151.7 |
| Qwen3.8 2.4T A95B | Aug 12, 2026 | $3 | — |
| Qwen3.7 Flash | Jul 27, 2026 | $0.06 | 144.1 |
| Qwen3.7 Plus | Jun 3, 2026 | $0.56 | 147 |
Verdict
Computed score: 40/100, from public data (performance, price, context, features).
Strengths
- Open weights: self-hostable, your data stays with you.
Limitations
- No tool calling: unusable as an agent's brain.
- No independent score published: performance cannot be verified.
- Older generation, often replaced by a newer model from the same vendor.
Avoid if…
- you are building an agent that must call tools.
Alternatives
Comparisons and tools
Learn more
Frequently asked questions
How much does Qwen2.5 VL 72B Instruct cost?
$0.8 per million input tokens and $1 per million output tokens on OpenRouter (checked October 25, 2026). A task with 2,000 tokens read and 500 written costs $0.0021.
How big is Qwen2.5 VL 72B Instruct's context window?
128,000 tokens, roughly 96,000 words, with a maximum of 115,200 output tokens.
Can Qwen2.5 VL 72B Instruct run locally?
Yes, its weights are published.
Is Qwen2.5 VL 72B Instruct suitable for an AI agent?
No: without tool calling it cannot act, only answer.
What are the alternatives to Qwen2.5 VL 72B Instruct?
The closest models on price and performance are Llama 3.1 Euryale 70B v2.2, Nano Banana (Gemini 2.5 Flash Image), and GLM 4.5V.
Related articles
- Mistral AI: models, pricing and agentsOctober 25, 2026
- LLM benchmarks: how to read model rankingsOctober 24, 2026
- Running a local LLM: hardware, software, and limitationsOctober 24, 2026
- Open-source or closed models for your AI agents?October 5, 2026
Neutrality. No affiliate links. No vendor pays to be reviewed, rated or ranked. Reviews are directed by François Chollet, neutral and factual, and every data point is dated and sourced. How we score