BitShovel
中

Qwen model card and HF trend record · First found 2026-09-05

Qwen3.8-27B: images, video and adjustable thinking

Qwen3.8-27B’s model card describes text, image and video understanding, adjustable thinking, open weights and deployment files.

Retained Qwen3.8-27B source cover from its official Hugging Face model page.
Image: Original image from Qwen’s model page · Open full-size image
In this articleContents3

中文

Input and thinking controls

The card describes a 27B-parameter language model with a vision encoder for text, image and video understanding. Thinking starts enabled and can be disabled or adjusted for each request.

Qwen3.8-27B combines text, image and video input with a vision encoder and per-request thinking controls.
Graphic 1/2: input modes and thinking controls from the model card checked on September 9. · Open full-size image
  • Model facts: 27.8B total params · license: apache-2.0 · data: HF repo metadata · observed 2026-10-25 09:03:06 Beijing time

中文

Context and available files

Native context is 262,144 tokens, expandable to 1,000,000. The repository distributes Transformers-format weights and configuration under Apache 2.0. Its official hosted version was still listed as coming soon when checked on September 9.

Native and extended Qwen3.8-27B context sizes, open weights and configuration, and the hosted version’s coming-soon status.
Graphic 2/2: context sizes, deployment files and the separately stated hosted-service status. · Open full-size image

中文

Dates and discovery

These specifications come from the September 9, 2026 model-card reading. The September 5 trend record’s “entry #2” is an order after BitShovel’s filters, not an original Hugging Face ranking or a feature-release date.

Sources and notes

sources中文

Sources and notes

Retained summary and context

Entry numbers describe this filtered collection, not original HF chart positions or global ranks. Trending score, downloads, and likes: attention is not proof of model quality, adoption, benchmark results, or production readiness. Check the official model page and model card.

Sources and evidence · 2 links

Each link has a purpose and scope. Link counts do not establish independent confirmation; direct support applies only to the statements specified below.

View the model data record

Source timeline
  1. · Entry #2 after this HF trend collection's filters, with a trending score of 551; no original HF chart rank is recorded here.
Collection record

First discovered:

Collection and source publication have distinct dates.

Editorial record

· Beijing time (UTC+8)

The title now names the image and video understanding and thinking controls described in the article, removing the internal collection order. The original trend snapshot and discovery dates remain; this title edit does not represent a new model release.

Sources and full record

This dates our review of the wording and sources, not a project release or product update.

Full record: before, after and sources (JSON)

Corrections and feedback