Model Releases
for some harder tests (this was insect identification), hosted models spread over 23 pts. Qwen 4B + Glance (0.96) was within a point of the …
On September 21 2026, Yohei reported results from a hard “insect‑identification” benchmark where model accuracy spanned 23 points across the evaluated systems. The Qwen 4B + Glance model scored 0.96,
On September 21 2026, Yohei reported results from a hard “insect‑identification” benchmark where model accuracy spanned 23 points across the evaluated systems. The Qwen 4B + Glance model scored 0.96, placing it within one point of the best performer and beating competitors such as Haiku 4.5 and GPT‑5.6. The results demonstrate a substantial performance gap among hosted models on this challenging test.
Related
- Evaluation Cards: An Interpretive Layer for AI Evaluation Reporting
- A Statistical Audit of Physical AI Benchmark Redundancy
Source: Yohei Nakajima (X) | 2026-09-21