Claude Opus 4.7
The most advanced publicly available language model, optimized for agentic coding, long-term autonomous tasks, and image reasoning.
Technical specification
Modalities
Capabilities
12Reasoning★
Reasoning
Multi-step reasoning★
Reasoning
Long context★
Reasoning
Coding★
Coding
Function Calling
Planning
Structured output★
Structured gen.
Image understanding★
Vision
Chart understanding
Vision
OCR★
Vision
Multilingual★
Language
Planning★
Planning
Streaming output
Reasoning
Standard API pricing: $5 per million input tokens, $25 per million output tokens. Prompt caching: cache write at $6.25/MTok (TTL 5 min) or $10/MTok (TTL 1 hr); cache read at $0.50/MTok. Batch API: 50% discount on input and output tokens (asynchronous processing, up to 24 hrs). Full 1M token context window at standard pricing — no surcharge for long context. In Batch API, up to 300K output tokens per request are available using the beta header output-300k-2026-03-24. US-only inference (inference_geo parameter) applies a 1.1× multiplier across all pricing tiers. Pricing unchanged from Claude Opus 4.6.
INPUT
$5.0000 / per 1M tokens
OUTPUT
$25.0000 / per 1M tokens
CACHE
$0.5000 / per 1M tokens
TOTAL
for 10K tokens
Synchronous Messages API. Cache write: $6.25/MTok (TTL 5 min) or $10/MTok (TTL 1 hr). Max 128K output tokens synchronously; up to 300K with the Batch API beta header.
Batch API (50% discount); asynchronous processing within 24 hours. Supports up to 300K output tokens per request using the beta header output-300k-2026-03-24.
Security and enterprise
Claude Opus 4.7 is the first Claude model with automated real-time cybersecurity safeguards that detect and block requests indicating prohibited or high-risk cyber use cases (Project Glasswing initiative). On Amazon Bedrock, a zero operator-access policy applies — customer prompts and responses are not visible to Anthropic or AWS employees. Data residency options are available via the inference_geo parameter in the Claude API (1.1× multiplier for US-only inference). Regional and multi-regional endpoints on Google Vertex AI and Amazon Bedrock are available at a 10% premium over global endpoints. Enterprise security and compliance information: trust.anthropic.com.
Technical information
Security professionals conducting legitimate work (penetration testing, red-teaming, vulnerability research) may apply for access through Anthropic's Cyber Verification Program. A full safety evaluation is published in the Claude Opus 4.7 system card (232 pages, April 16, 2026). The compliance assessment concluded that the model is "largely well-aligned and trustworthy, though not fully ideal."