1M Context Is Now Generally Available for Opus 4.6 and Sonnet 4.6

Why It Matters

Providing a true 1 M token context at standard rates removes a major cost and engineering barrier, unlocking richer AI applications for enterprises and developers.

Key Takeaways

• 1M context now standard pricing, no long‑context premium
• Opus 4.6 priced $5 per million input, $25 output
• Sonnet 4.6 priced $3 per million input, $15 output
• Media limit rises to 600 images or PDF pages
• Beta header ignored; no code changes required for >200K tokens

Pulse Analysis

The move to a universal 1 million‑token context window marks a watershed for large‑language‑model deployments. Previously, developers faced steep premiums or complex token‑management logic to handle extensive documents, codebases, or multi‑modal data. By flattening the pricing structure—$5/$25 per million tokens for Opus 4.6 and $3/$15 for Sonnet 4.6—Claude eliminates a hidden cost layer, making high‑capacity reasoning financially predictable for enterprises scaling AI workloads.

Beyond pricing, the sixfold increase in media capacity to 600 images or PDF pages reshapes how organizations embed rich, multimodal content into prompts. Legal teams can now feed full contracts, scientists can process hundreds of research figures, and software engineers can analyze massive code diffs without chunking. The removal of the beta header requirement further simplifies integration, allowing existing pipelines to automatically benefit from the expanded context without code revisions. This operational ease accelerates time‑to‑value for AI‑augmented products across cloud providers like Azure Foundry and Google Vertex AI.

Performance gains reinforce the business case: Opus 4.6 achieves a 78.3 % MRCR v2 score, the highest among frontier models at this scale, ensuring that the larger context does not dilute accuracy. Real‑world customers report up to 15 % fewer compaction events and smoother agent workflows, translating into higher‑quality outputs and lower token waste. As enterprises increasingly rely on AI for complex decision‑making—ranging from contract negotiation to incident response—the ability to retain a full conversation history without manual summarization becomes a competitive differentiator, positioning Claude as a go‑to platform for deep‑context AI applications.