In a rapid succession of releases, xAI launched Grok 4.7, Anthropic released Claude Opus 5.5, and OpenAI introduced two GPT-6 models, Sol and Luna, all within a 48-hour period. This flurry of activity has rendered previous pricing models obsolete, shifting the competitive focus from per-token costs to per-task expenses. While Opus 5.5 offers a 20% cost reduction and improved performance, OpenAI's new GPT-6 models, Sol and Luna, are highlighted for their significantly lower per-task costs, especially in complex reasoning scenarios, though this efficiency may come at the expense of initial response latency. AI
IMPACT Shifts focus from token pricing to task-based costs, requiring users to re-evaluate AI model expenses based on actual workload efficiency.
RANK_REASON Multiple frontier labs (xAI, Anthropic, OpenAI) announced new model versions (Grok 4.7, Claude Opus 5.5, GPT-6 Sol/Luna) within a short timeframe.
- Agents’ Last Exam v1
- Anthropic
- Artificial Analysis
- Astra
- Claude Opus-5
- Claude Opus 5.5
- DeepSWE 1.1
- GPT-5.6
- GPT-6
- Grok 4.7
- Luna
- .openai
- OSWorld 2.0
- Sol
AI-generated summary · Google Gemini · from 5 sources. How we write summaries →