Qwen2.5-VL-7B-Instruct
- In Input modalities: TextImage Out Output modalities: Text
- Open weights
Prices
2 offerings · prices as listedqwen2.5-vl-7b-instruct| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard | $0.35 | $1.05 |
Prices from models.dev: qwen2-5-vl-7b-instruct (community data, MIT) (opens in a new tab) (checked Oct 15, 2026)
qwen2.5-vl-7b-instruct| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard | ¥2.00 | ¥5.00 |
Prices from How to check model training and deployment pricing-Alibaba Cloud Model Studio(Model Studio)-阿里云帮助中心 (opens in a new tab) (checked Nov 1, 2026)
Standard: the price a request with no conditions pays (the headline price elsewhere on the site).
Details
- Reasoning
- —
- Tool use
- —
- Structured output
- —
- Knowledge cutoff
- Apr 2024
- Licence
- apache-2.0
- Family
- qwen-vl
- Model ID
Not stated on Alibaba Qwen’s pages: max input, reasoning efforts, reasoning can be off, max thinking budget, built-in tools, aliases.
Run locally
Weights: Qwen/Qwen2.5-VL-7B-Instruct- Parameters 8.3B
- Dense
- Native precision BF16
- Layers 28
- KV cache per token 56 KB
- Config positions 125K (config.json max_position_embeddings)
- API context 977K (maker’s API)
- Needs
- ≈ 16.7 GB (weights 15.4 (maker’s file) + KV cache 0.4 + overhead 0.8)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 30–44 tok/s
- Longest context
- 98K (longer doesn’t fit)
- First token
- about 0.27–0.40 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 15.4 GB | ≈ 16.7 | ●Runs well 30–44 tok/s | 98K |
| Q8_0 | 8.2 GB | ≈ 9.1 | ●Runs well 54–81 tok/s | 125K |
| FP8 | 7.7 GB | ≈ 8.7 | ●Runs well 58–86 tok/s | 125K |
| Q6_K | 6.3 GB | ≈ 7.3 | ●Runs well 69–104 tok/s | 125K |
| Q5_K_M | 5.5 GB | ≈ 6.4 | ●Runs well 79–119 tok/s | 125K |
| Q5_0 | 5.3 GB | ≈ 6.2 | ●Runs well 82–123 tok/s | 125K |
| Q4_K_M | 4.7 GB | ≈ 5.6 | ●Runs well 92–137 tok/s | 125K |
| Q4_0 | 4.3 GB | ≈ 5.3 | ●Runs well 98–147 tok/s | 125K |
| INT4 (AWQ / GPTQ)maker’s file | 6.4 GB | ≈ 7.4 | ●Runs well 68–102 tok/s | 125K |
| IQ4_XS | 4.1 GB | ≈ 5.0 | ●Runs well 103–155 tok/s | 125K |
| Q3_K_Mquality drops | 3.8 GB | ≈ 4.7 | ●Runs well 111–167 tok/s | 125K |
| IQ3_XXSquality drops | 3.0 GB | ≈ 3.9 | ●Runs well 138–208 tok/s | 125K |
| Q2_Kquality drops | 2.9 GB | ≈ 3.8 | ●Runs well 141–211 tok/s | 125K |
| IQ2_XXSquality drops | 2.0 GB | ≈ 2.9 | ●Runs well 193–290 tok/s | 125K |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU GeForce RTX 3080; unified memory Apple M5 · 16 GB.
Run commands
vllm serve Qwen/Qwen2.5-VL-7B-Instruct --max-model-len 8192
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated
Cited scores
From third-party leaderboards, not measured by usThis model isn’t on any leaderboard we cite yet.
Changes
All changes →- Now priced on Alibaba Cloud Model Studio, China Input ¥2.00 · Output ¥5.00
- Now priced on Alibaba Cloud Model Studio, International Input $0.35 · Output $1.05
- New language model
Sources
- Text generation - Alibaba Cloud Model Studio (opens in a new tab) docs.modelstudio.console.alibabacloud.com · checked Oct 15, 2026
- Text generation models - Qwen, DeepSeek, GLM | Model Studio - Alibaba Cloud Model Studio - Alibaba Cloud Documentation Center (opens in a new tab) docs.modelstudio.console.alibabacloud.com · checked Oct 15, 2026
- models.dev: qwen2-5-vl-7b-instruct (community data, MIT) (opens in a new tab) github.com · checked Oct 15, 2026
- Qwen/Qwen2.5-VL-7B-Instruct model card (opens in a new tab) huggingface.co · checked Nov 1, 2026
- Qwen/Qwen2.5-VL-7B-Instruct config.json (opens in a new tab) huggingface.co · checked Oct 31, 2026
- Qwen/Qwen2.5-VL-7B-Instruct files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- Qwen/Qwen2.5-VL-7B-Instruct-AWQ files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- Qwen/Qwen2.5-VL-7B-Instruct model card (opens in a new tab) huggingface.co · checked Nov 1, 2026
- How to check model training and deployment pricing-Alibaba Cloud Model Studio(Model Studio)-阿里云帮助中心in Chinese (opens in a new tab) help.aliyun.com · checked Nov 1, 2026
Alibaba Qwen’s official policies: data use, privacy, terms →
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Alibaba Qwen · China · qwen.ai
Related models
Snapshots, the same family and other models from this provider