QwenReactBench

QwenReactBench - in-house React project building benchmark using Claude Code harness. Bilingual (EN/CN), 7 categories; auto-render + multimodal judge; BT/Elo rating.

Category

Agentic Coding

Max Score

—

Score Type

elo

Active

Yes

Model Scores

Scores for QwenReactBench

Model Score Date Verified
Fable-5
100.00%
18.08.2026 Unverified
GPT-5.6 Sol
11.21%
18.08.2026 Unverified
Claude Opus 4.8
67.24%
18.08.2026 Unverified
Qwen3.7-Max
1538.00
18.08.2026 Unverified
Qwen3.8-2.4T-A95B
80.17%
18.08.2026 Verified