1M context is now generally available for Opus 4.6 and Sonnet 4.6
Anthropic's Claude Opus 4.6 and Sonnet 4.6 now boast a 1 million token context window, available at standard pricing with no premium, a significant leap for large-scale AI applications. This move streamlines development by eliminating complex context management and extra costs, sparking intense discussion among Hacker News users about its real-world performance, billing implications, and competitive positioning against other frontier models. The community is eagerly testing whether such a vast context window truly maintains model coherence or if degradation still occurs at scale.
The Lowdown
Anthropic has announced the general availability of a 1 million token context window for its Claude Opus 4.6 and Sonnet 4.6 models. This significant upgrade comes without any long-context premium, making advanced AI capabilities more accessible and cost-effective for developers.
- Standard Pricing: The full 1M token window is now billed at standard per-token rates (e.g., $5/$25 per million tokens for Opus), eliminating previous surcharges for extended context.
- Enhanced Media Limits: Users can now process up to 600 images or PDF pages per request, a sixfold increase from the previous limit.
- Seamless Integration: The 1M context is available automatically, no longer requiring a beta header, simplifying implementation.
- Automated for Claude Code: Max, Team, and Enterprise users of Claude Code with Opus 4.6 will automatically benefit from the full 1M context, leading to fewer compactions and more persistent conversations.
- Proven Coherence: Anthropic claims Opus 4.6 achieves a 78.3% score on MRCR v2, demonstrating strong recall and reasoning across the extensive context window.
- Real-World Applications: This expanded capacity enables handling entire codebases, thousands of pages of legal documents, or comprehensive agent traces in a single session, reducing the need for lossy summarization and complex engineering work.
This update marks a substantial advancement for large language models, promising to transform workflows requiring deep contextual understanding and continuous interaction with vast amounts of information by making it more affordable and efficient.
The Gossip
Pricing & Premium Practicality
Commenters are thrilled that Anthropic has removed the 'long-context premium,' making the 1M token window available at standard pricing. This is seen as a major competitive play, potentially putting pressure on rivals like OpenAI who might still charge extra for extended context. Users highlight the financial implications, suggesting it will significantly lower costs for large-scale applications.
Effective Contextual Coherence
A recurring concern among users is the practical efficacy of such a large context window. While impressive on paper, many have experienced performance degradation or 'context rotting' in previous models as the context length approached its stated limit. The discussion revolves around whether Claude truly maintains reasoning ability and recall across the full 1M tokens, or if it's merely a larger bucket that still suffers from diminishing returns.
Deployment & Workflow Shifts
Users are noting immediate, sometimes unexpected, changes in their AI workflows due to the automatic rollout of the 1M context. While generally welcomed as a powerful enhancement, some express a slight unease about unannounced behavior changes in deployed models. Others are excited about the practical implications, such as fitting entire coding sessions or complex debugging processes into a single, continuous AI interaction, hinting at a significant shift in how they develop and interact with these models.