Model evidence

Qwen 3.8

Read the measured Qwen 3.8 Max configuration without treating missing cost or an embedded Max label as a universal effort scale.

Current batch

Qwen 3.8 / Max is both the highest-scoring measured configuration at 66/100 and the fastest in this family at 21m 19s in the selected batch.

Open the complete Radar Read the evaluation method

Selected evidence

Published: Aug 26, 2026, 11:00 PM UTC
Question pack: coding-fast-v4.10
Coverage: 1 measured configuration
Batch: snapshot-2026-08-26T23-00-00Z-r3

Measured configurations

Configuration	Rank	Score	Elapsed	Reference cost	Route
Qwen 3.8 / Max	#10	66/100	21m 19s	$0.89	Custom endpoint

Five-question profile

Question	Qwen 3.8 / Max
Black-box Regression Audit	10/20
Retry Planner Counterexamples	17/20
CI Adversarial Audit	12/20
Transaction Regression Design	13/20
Cache Regression Test Design	14/20

Why the raw configuration matters

The measured tier is encoded in the Qwen model identity while the endpoint can expose a default effort field. ModelDial presents it as Max but retains the raw configuration so the result stays reproducible.

If only one Qwen 3.8 configuration is present in the selected batch, the page can compare it with other configurations but cannot infer a Qwen effort curve.