AI
πCursor Launches Composer 2: The $0.50 Frontier Model That Could Reshape Coding
The Rundown: Cursor released Composer 2, a reinforcement learning-trained coding model scoring 61.3 on CursorBench at $0.50/M input tokens, directly competing with OpenAI and Anthropic.
The details:
- βCursor Composer 2 achieves 61.3 on CursorBench using reinforcement learning on long-horizon multi-file coding tasks, priced competitively at $0.50/M input tokens
- βGoogle AI Studio now generates full-stack apps with frontend, backend, and Firebase integration from natural language prompts via the Antigravity agent
- βAnthropic conducted AI usage research with 80,508 users across 159 countries, finding 81% report AI helped them reach goals with productivity as the top use case
Why it matters: This signals the beginning of vertical AI model competition where coding-specific models could outperform general-purpose LLMs at significantly lower costs. For founders, this means dramatically reduced development costs and the potential for smaller teams to build complex software systems.
π° Source: AlphaSignal