NutsNews summary screening: Qwen 2.5 3B versus 7B

Historical screening results and their production limitations.

Visual overview

Primary diagram

System map

A small controlled model screening provides observations, but production stays unchanged until representative repeated chat tests and human review support a change.

Render the repository-owned system map when you need it.

Diagram is not rendered yet.

View as text

NutsNews summary screening: Qwen 2.5 3B versus 7B decision

A small controlled model screening provides observations, but production stays unchanged until representative repeated chat tests and human review support a change.

flowchart TD
  accTitle: NutsNews summary screening: Qwen 2.5 3B versus 7B decision
  accDescr: A small controlled model screening provides observations, but production stays unchanged until representative repeated chat tests and human review support a change.
  inputs["Three source-checked story digests"] --> test["One controlled run per story and model"]
  test --> results["Record speed, length and factual observations"]
  results --> limits["Small sample; not the production chat workflow"]
  limits --> decision["Keep production unchanged"]
  decision --> future["Repeated representative chat tests and human review"]

NutsNews summary screening: Qwen 2.5 3B versus 7B

Three story digests were tested once with Qwen 3B and 7B on September 13, 2026. This used a controlled generation test, not the exact production chat workflow. Production stayed on its existing model.

3B produced about 14.7 tokens per second, compared with about 7.0 for 7B. All three 3B summaries met the 200–250 character limit; none of the 7B summaries did. The small manual review found factual weaknesses in both models and no clear overall advantage for 7B.

Health requests succeeded during the test, but that does not prove unchanged publishing speed or AI queue latency. The test model was unloaded afterward. Keep production unchanged, improve summary validation first, and require repeated production-chat tests and human review before considering a model switch.