Model · Qwen

Qwen Plus qwen3.7

4 appearances · meta-score 78.5/100 · last seen 2026-07-22 · scored under rubric v15

Pooled composite across 4 primary runs: 68.9 ± 18.5 (95% CI)

$100.79Total cost
522,162,473Tokens
67.8%Cache hit
7,049Requests
0.03%Failure rate
1.14Tools/response
5,180,471Tokens per $

Agent role scores

Per-agent-role scores for Qwen Plus qwen3.7
Agent role	Score / 10
Code-Review Agent	8.84
Sprint-Review Agent	8.81
Architect	8.49
Orchestrator	8.45
Develop Agent	8.15
Sprint-Plan Agent	7.63
Build Agent	7.59
Director	7.43
Pseudocode Agent	7.38
Code Agent	6.97
Test Agent	6.92
Platform Agent	6.59
Scout Agent	6.30

Run history (primary model)

Runs where Qwen Plus qwen3.7 was the primary model, oldest first
Run	SoW	Date	Composite
Plus qwen3.7	circles	Jul 2026	64.3
Plus qwen3.7	circles	Jul 2026	79.9
Plus qwen3.7	solarSystem	Jul 2026	54.7
Plus qwen3.7	circles	Jul 2026	76.7

Every run on Favur Evals is scored by the same deterministic engine, on the same Statements of Work. The benchmark is self-funded — no vendor sponsorship, credits, or grants.