# Qwen3 VL 8B Thinking — price & specs

Source: https://modelbank.ai/models/qwen3-vl-8b-thinking · updated 2026-11-12 · ModelBank

| Field | Value |
|---|---|
| Vendor | Qwen (Alibaba) |
| Input price | $0.18 / 1M tokens |
| Output price | $2.1 / 1M tokens |
| Cached input | n/a / 1M tokens |
| Context window | 131,072 tokens |
| Max output | 32,768 tokens |
| Input modalities | image,text |
| Reasoning / tool use | yes / yes |
| Intelligence / coding index | 8.2 / — |
| Released | 2025-10-14 |
| Retirement date | 2026-10-09 |

## About

Qwen3-VL-8B-Thinking is the reasoning-optimized variant of the Qwen3-VL-8B multimodal model, designed for advanced visual and textual reasoning across complex scenes, documents, and temporal sequences. It integrates enhanced multimodal alignment and...

## Price history

- 2026-09-24: $0.18 in / $2.1 out

## Cheaper alternatives with similar intelligence

- [GLM 5.3 Flash](https://modelbank.ai/models/glm-5.3-flash): $0.0399 / $0.0784, intelligence 27.8
- [Nemotron 3.5 Lightning](https://modelbank.ai/models/nemotron-3.5-lightning): $0.07 / $0.2, intelligence 5.5
- [DeepSeek V4.1 Flash 1017](https://modelbank.ai/models/deepseek-v4.1-flash-1017): $0.098 / $0.29, intelligence 28.6
- [MiMo-V2.6-Flash](https://modelbank.ai/models/mimo-v2.6-flash): $0.14 / $0.28, intelligence 26.7
- [Claude Haiku 5.5](https://modelbank.ai/models/claude-haiku-5.5): $0.1 / $0.5, intelligence 36.5

JSON: https://modelbank.ai/api/v1/models/qwen3-vl-8b-thinking