← All models

hunyuan-turbos-vision-20250619

Published Publisher Tencent

Positioning and release date

The 2025-06-19 version of Tencent's Hunyuan TurboS Vision series of flagship vision-language models, intended for image-text understanding tasks. Tencent Cloud product updates record its launch date as June 19, 2025. (cloud.tencent.com)

Capabilities and uses

Tasks described officially include image-based entity recognition, knowledge question answering, copywriting, and solving problems from photographs. Tencent states that performance on these image-text understanding tasks has improved across the board compared with the previous-generation default version; the announcement does not provide quantitative metrics for the improvements in this version. (cloud.tencent.com)

Limitations

  • The release announcement does not explicitly list this fixed version's context window, image count and size limits, maximum output length, or supported languages. Specific limits are subject to the actual service documentation and invocation configuration.
  • Image understanding and problem-solving results may be incorrect; users should verify important content.
  • The announcement concerns image-text understanding; the capabilities of a separately released video understanding model should not be attributed directly to this version. (cloud.tencent.com)

Pricing

The list prices for the Hunyuan-turbos-vision series in the Tencent Cloud China site's Hunyuan text generation pay-as-you-go price list are: input 3 CNY / million tokens, output 9 CNY / million tokens. The price list is organized by series name and does not separately identify the fixed 20250619 version, so the specific applicable pricing for this version must still be confirmed when enabling or invoking the service; free allowances and resource packages may affect actual billing. Official pricing source: Hunyuan Text Generation Billing Overview, consulted on: 2026-11-18. (cloud.tencent.com)

Official sources