WebCoderBench: leaderboard
Metric: Overall score: preference-weighted sum of metric z-scores (% of a standard deviation). Source: arxiv.org. Saturation forecast: Around January 2027. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 (High) | 13.81 |
| 2 | GLM-4.5 | 11.76 |
| 3 | MiniMax-M2 | 7.89 |
| 4 | DeepSeek V3.1 Terminus | 6.26 |
| 5 | DeepSeek R1 0528 | 2.41 |
| 6 | Gemini 2.5 Pro | 0.27 |
| 7 | GPT-5 Codex (High) | -0.03 |
| 8 | Qwen 3 235B A22B Instruct | -0.04 |
| 9 | Qwen 3 Coder Plus | -2.31 |
| 10 | DeepSeek V3.1 (Thinking) | -3.73 |
| 11 | DeepSeek V3.1 | -3.88 |
| 12 | GPT-4o (2024-11-20) | -38.65 |
Interactive version: theaggregate.ai/benchmark?slug=webcoderbench · How It Works · Data refreshed daily, snapshot 2026-09-25.