Anthropic leads the LiveBench Coding market at 86.5% implied probability because its Claude Opus 5.5 and Fable 5 models have topped recent coding snapshots, including the September 2026 resolution where they posted measurable margins over competitors on contamination-resistant tasks. Traders see continued strength in code generation, debugging, and agentic workflows relative to OpenAI’s GPT-6 Astra, which sits at 9.6% and trails on several coding sub-benchmarks. Google, Meta, and other labs remain at single-digit or lower odds amid weaker recent coding results. With the October window still open, any new model release or sharp leaderboard shift could alter positioning before month-end resolution.