Qwen 3.8
Model version · Mentioned in posts · release unverified
21 collected posts from 18 authors.
Published snapshot: · Checking for updates…
We collected 58 documents mentioning Qwen over seven days. The reading below describes that sample of discussion, not the views of everyone who uses it.
Lower = more trust-leaning discussion; higher = more concern. This is sentiment, not a test of reliability or safety.
Small differences near 50 are not evidence of a better model. Read the original sources below.
Announcements, owner-tracked versions and reactions collected over the last 7 days. Devices and assistants are labelled separately from models.
Model version · Mentioned in posts · release unverified
21 collected posts from 18 authors.
Each full update checks new version names. Automatically discovered names need several authors; owner-tracked names are labelled separately. Neither gets a ranking without enough evidence.
Original · EN · English · detectedIf this holds up, it is big: "We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training." rikiyat.github.io/blog/weight-...
Linked origin: bsky.app.
No publisher-policy review is recorded here. Unknown does not mean unreliable or fake.
Claim: not fact-checked by this project. Sentiment analysis is not fact-checking.
Account identity and expertise are not independently verified. Likes, followers and reposts measure attention, not truth.
Emotional scores are machine-generated estimates about a statement, not an assessment of the speaker. Visible excerpts have identified source languages and cached English and Hungarian translations. Untranslated or unknown-language excerpts are held back, but may still contribute to the index.