OCR and document images
Extract and reason over text in screenshots, photographed pages, forms, diagrams, and other image-based documents.
Independent tool overview
Hunyuan Vision 1.5 Thinking at a glance
Hunyuan Vision 1.5 Thinking is Tencent's deep-reasoning vision-language model for image question answering, visual grounding, OCR, charts, STEM problems, and image-based creative tasks. It remains available through Tencent Cloud TokenHub under the model ID hunyuan-t1-vision-20250916.
Visit the official Hunyuan Vision 1.5 Thinking site ↗
- Developer
- Tencent Hunyuan
- Model ID
- hunyuan-t1-vision-20250916
- Access
- Tencent Cloud TokenHub
- Context window
- 40K tokens
- Maximum output
- 24K tokens
- Media support
- One image per request; no video
Overview
What Hunyuan Vision 1.5 Thinking is
Hunyuan Vision 1.5 Thinking analyzes an image together with a text instruction and produces a reasoned text response. Its intended workloads include locating objects, reading text, interpreting diagrams and charts, solving photographed problems, and answering questions that require several visual reasoning steps.
Tencent describes the model as a Mamba-Transformer hybrid with a thinking-on-images approach. The research framing includes visual reflection operations such as cropping, zooming, and drawing points, lines, or boxes to inspect an image more deliberately.
The serving path changed in 2026. Tencent retired the old Hunyuan multimodal endpoint on June 22 and moved HY-Vision-1.5-Thinking to TokenHub, which exposes an OpenAI-compatible API. The model is active, but integrations must use the new endpoint.
Use cases
Who Hunyuan Vision 1.5 Thinking is best for
The strongest fit depends on the job you need the product to complete, not the size of its feature list.
Charts and visual data
Interpret chart structure, compare plotted values, explain visual patterns, and answer questions grounded in an image.
STEM problem solving
Analyze photographed exercises, geometry, formulas, and other visual problems that benefit from deep reasoning.
Multilingual applications
Build image-understanding workflows for Chinese and more than 30 listed languages through Tencent TokenHub.
Capabilities
Core Hunyuan Vision 1.5 Thinking features
Deep visual reasoning
Uses a thinking-oriented response path for questions that require several reasoning steps instead of only a short caption.
Visual grounding
Identifies and reasons about where an object or relevant region appears within an input image.
OCR and chart understanding
Reads visible text and interprets structured visual information such as tables, plots, diagrams, and screenshots.
Image question answering
Answers general questions, analyzes scene content, and supports multi-turn conversations grounded in an uploaded image.
Multilingual interaction
Tencent lists more than 30 major languages and highlights improved English and smaller-language performance.
OpenAI-compatible API
Calls the model through TokenHub's chat-completions-compatible endpoint using the OpenAI SDK or another compatible client.
Process
How the Hunyuan Vision 1.5 Thinking workflow works
Step 1
Create a TokenHub service
Use Tencent Cloud TokenHub rather than the retired Hunyuan endpoint, then enable the model's free trial or pay-as-you-go service.
Step 2
Configure the client
Point an OpenAI-compatible SDK to the TokenHub base URL and select hunyuan-t1-vision-20250916 as the model.
Step 3
Prepare one clear image
Crop to the relevant content and preserve readable resolution because HY-Vision accepts one image per request.
Step 4
Ask a grounded question
Specify the OCR, localization, comparison, chart, STEM, or explanation task and request evidence tied to visible regions.
Step 5
Validate the response
Check extracted text, labels, units, coordinates, calculations, and conclusions against the source image.
Step 6
Monitor the integration
Track token usage, confirm the TokenHub endpoint in every environment, and compare HY-Vision-2.0-Instruct when speed matters more than deep reasoning.
Cost
Hunyuan Vision 1.5 Thinking pricing and free plan
TokenHub bills Hunyuan Vision 1.5 Thinking by input and output tokens. Tencent lists ¥3 per million input tokens and ¥9 per million output tokens. New TokenHub accounts can claim a one-million-token multimodal trial allowance valid for one year; the current trial campaign is listed through December 31, 2026.
New-user trial
1M tokens free
One-time TokenHub trial allowance for multimodal-understanding models.
- Valid for one year after claiming
- Available once per primary account
- Current campaign listed through December 31, 2026
Pay-as-you-go input
¥3 per 1M tokens
Postpaid price for prompt text and encoded image input.
- Billed hourly
- Free allowance is consumed first
- Postpaid billing must be enabled after the trial
Pay-as-you-go output
¥9 per 1M tokens
Postpaid price for generated response tokens.
- Long reasoning uses more output tokens
- Usage is visible in TokenHub
- Prices are in Chinese yuan
Pricing checked . Check current pricing at the source ↗
Assessment
Hunyuan Vision 1.5 Thinking strengths and limitations
Where it stands out
- Deep reasoning for image-grounded questions
- Covers OCR, charts, visual localization, STEM problems, and image analysis
- Supports more than 30 listed languages
- OpenAI-compatible endpoint reduces client integration work
- 40K context window and up to 24K output tokens
- Low published token prices and a new-user trial
What to consider
- Accepts one image per request and does not support video input
- The old Hunyuan multimodal endpoint was retired, so older integrations require migration
- A newer HY-Vision-2.0-Instruct model is available for fast-thinking image tasks
- The GitHub repository still marks the technical report and downloadable checkpoints as forthcoming
- Launch benchmark claims are creator-reported and date to 2025
- OCR, chart values, spatial claims, and reasoning steps need source-image verification
Compare
Hunyuan Vision 1.5 Thinking alternatives
The right alternative depends on the specific output, workflow, controls and budget your project requires.
Consumer
Gemini 3.1 Pro
Choose Gemini 3.1 Pro when a newer broadly multimodal flagship and Google's developer ecosystem are a better fit.
Explore Gemini 3.1 Pro →Consumer
Qwen3.5-Omni
Choose Qwen 3.5 Omni when one model must understand text, images, audio, and video across a broader language set.
Explore Qwen3.5-Omni →Project Management
Claude
Choose Claude when document-heavy visual analysis and Anthropic's general assistant workflow matter more than Tencent deployment.
Explore Claude →Questions
Hunyuan Vision 1.5 Thinking FAQs
What is Hunyuan Vision 1.5 Thinking?
It is Tencent's deep-reasoning vision-language model for image question answering, visual grounding, OCR, charts, photographed problems, and image-based creative analysis.
Is Hunyuan Vision 1.5 Thinking still available?
Yes. The old Hunyuan multimodal service ended in June 2026, but Tencent moved the model to TokenHub, where it remains listed under hunyuan-t1-vision-20250916.
How much does Hunyuan Vision 1.5 Thinking cost?
Tencent TokenHub currently lists ¥3 per million input tokens and ¥9 per million output tokens. Eligible new accounts can claim a one-million-token trial allowance valid for one year.
Can it analyze video?
No. Tencent lists image understanding but not video understanding for this model. HY-Vision-Video and YT-VITA are the relevant Tencent options for video inputs.
Is Hunyuan Vision 1.5 Thinking open source?
Not currently. Tencent announced an open-source plan in 2025, but the official repository still shows the report and checkpoints as forthcoming. The usable release is a hosted TokenHub API.
Bottom line
Our Hunyuan Vision 1.5 Thinking verdict
Hunyuan Vision 1.5 Thinking is a practical, inexpensive choice for Tencent Cloud users who need deliberate OCR, chart, STEM, and image-grounded reasoning. Compare it with HY-Vision-2.0-Instruct and broader multimodal models when multiple images, video, or global platform access are required.
Visit Hunyuan Vision 1.5 Thinking website ↗Similar Miscellaneous Tools
View all Miscellaneous AI tools →
Muse Voice Transcribe
Muse Voice Transcribe - Meta's live speech-to-text that tracks 20+ speakers and mid-sentence language switches
Fish Audio S1
Fish Audio S1 -
Gemini Omni 1.1 Flash
Google's video model update with scene extensions and 4K upscaling
PokeeResearch 7B
PokeeResearch 7B- Pokee AI's SOTA, open-source deep research agent
AI training for the future of work.
Get access to all our AI courses, hundreds of real-world AI use cases, live expert-led workshops, an exclusive network of AI early adopters, and more.
AI Courses
Get unlimited access to all of our current & upcoming industry-specific AI courses for the duration of your subscription.
Daily Guides
To keep up with the rapid pace of AI, our team publishes AI implementation guides daily. Our library contains 300+ practical use cases to automate real-world work.
Workshops
Join weekly, live, interactive sessions with industry leaders who are at the forefront of AI for hands-on implementation guidance and exclusive insights.
Community
Network with an exclusive community of AI-first professionals who are working smarter with AI. Learn how early adopters are using AI in their work and businesses.