**Competitive parity among mid-tier labs is sustaining elevated uncertainty around third place on the Code Arena WebDev leaderboard.** As of late September 2026, Anthropic leads with Claude Opus 5.5 Max near 1820 Elo, followed closely by OpenAI’s GPT-6 Astra, while Alibaba’s Qwen3.8 Max sits third around 1671, just ahead of Moonshot’s Kimi K3 Max and Meta models in the mid-1650s. The tight clustering of scores from Alibaba, Moonshot, Meta, SpaceXAI, Tencent, and Z.ai reflects rapid iteration on agentic, multi-step frontend coding tasks, where cost-efficient Chinese open-weight releases have narrowed gaps with proprietary frontier models. Traders appear to view the next three months as wide open, with any major model update or benchmark surge capable of reordering the pack before year-end resolution.
基於Polymarket數據的AI實驗性摘要。這不是交易建議,也不影響該市場的結算方式。 · 更新於View resolved

















警惕外部連結哦。
警惕外部連結哦。
Frequently Asked Questions