Anthropic's Claude Fable 5.1 and Opus 5.5 models lead the latest LiveBench release with overall scores of 88.0 and 86.5, respectively, outpacing OpenAI's GPT-6 variants and Google's Gemini 4 series by several points across reasoning, coding, and agentic tasks. This recent performance on the contamination-resistant benchmark has driven trader consensus to 94% implied probability for Anthropic by end of October. OpenAI and Google trail despite competitive sub-benchmark results, while Chinese labs like DeepSeek remain several points behind. A new frontier model drop or LiveBench update in the next two weeks could still narrow the gap before resolution.
Экспериментальная сводка, созданная ИИ на основе данных Polymarket. Это не является торговой рекомендацией и не влияет на то, как разрешается этот рынок. · ОбновленоView resolved






















Не доверяй внешним ссылкам.
Не доверяй внешним ссылкам.
Часто задаваемые вопросы