Anthropic’s Opus 4.8 Tops Gemini 3.1 Pro; Google Engineer Charged Over $2.75M Bets
Anthropic released Claude Opus 4.8, claiming benchmark wins over Google’s Gemini 3.1 Pro and GPT-5.5 with agentic coding and financial analysis enhancements, priced at $5/$25 per million tokens (fast at $10/$50). Separately, DOJ charged a Google engineer for $2.75M in Polymarket bets using nonpublic Year in Search data, netting $1.2M.
1. Anthropic Launches Claude Opus 4.8
Anthropic unveiled its flagship Claude Opus 4.8 model as an upgrade to Opus 4.7, adding dynamic workflows and user effort controls for task tuning across browser and Claude Cowork interfaces.
2. Opus 4.8 Tops AI Benchmarks and Matches Pricing
The model outperformed OpenAI’s GPT-5.5 and Google’s Gemini 3.1 Pro in synthetic agentic coding, financial analysis, and computer use benchmarks, and maintains Opus 4.7’s pricing at $5 per million input tokens and $25 per million output tokens ($10/$50 in fast mode).
3. Google Engineer Faces Insider Trading Charges
A staff software engineer allegedly placed $2.75M in Polymarket bets based on confidential Year in Search data, netting $1.2M; federal prosecutors charged him with commodities and wire fraud and money laundering, and the CFTC filed a parallel civil suit seeking penalties and bans.
