Claude Opus 4.7 Analysis and Comparison

Claude Opus 4.7 Analysis and Comparison | Generated by AI

Home 2026.04

Question: How is Opus 4.7? Is it good? What’s the feedback? What’s different from Opus 4.6?

Answer:

I need to search for information about Claude Opus 4.7 to provide you with accurate details about this model, its performance, feedback, and differences from Opus 4.6.

Let me search for this information.Based on my search, I now have comprehensive information about Claude Opus 4.7. Let me provide you with a detailed answer.

Question: How is Opus 4.7? Is it good? What’s the feedback? What’s different from Opus 4.6?

Answer:

Claude Opus 4.7 was just released today (April 16, 2026) and represents a notable improvement over Opus 4.6, particularly in advanced software engineering and complex coding tasks.

Overall Assessment

Opus 4.7 is a notable improvement on Opus 4.6 in advanced software engineering, with particular gains on the most difficult tasks. Users report being able to delegate their hardest coding work to Opus 4.7 with confidence, including tasks that previously needed close supervision. The model handles complex, long-running tasks with greater rigor and consistency.

Key Improvements Over Opus 4.6

1. Advanced Coding Capabilities

Multiple companies reported significant improvements in coding benchmarks
On a 93-task coding benchmark, Claude Opus 4.7 lifted resolution by 13% over Opus 4.6, including four tasks neither Opus 4.6 nor Sonnet 4.6 could solve
Better at finding and fixing bugs, with improved code quality and fewer errors
More capable of handling multi-step workflows and long-running autonomous tasks

2. Enhanced Vision Capabilities

It can see images in greater resolution, processing images up to 2,576 pixels on the long edge (~3.75 megapixels), more than three times as many as prior Claude models
Dramatically improved visual acuity in testing (98.5% on one benchmark versus 54.5% for Opus 4.6)

3. Better Instruction Following

Opus 4.7 is substantially better at following instructions
Takes instructions more literally and precisely, though this means prompts written for earlier models may need adjustment

4. Improved Professional Output Quality

More tasteful and creative when completing professional tasks
Produces higher-quality interfaces, slides, and documents
Better design decisions and stronger creative choices

5. Enhanced Finance and Knowledge Work

State-of-the-art performance on finance agent evaluations
Opus 4.7 scored higher than its predecessor on benchmarks including finance agent evaluations and GDPval-AA, which measures economically valuable knowledge work across finance and legal domains

User Feedback

The feedback from early access testers has been overwhelmingly positive across various industries:

Coding/Development: Companies like Cursor, Replit, Cognition (Devin), and others reported double-digit improvements in task completion rates
Design: Users noted it’s the best model for building dashboards and data-rich interfaces with genuinely surprising design taste
Legal: Achieved 90.9% accuracy on BigLaw Bench at high effort
Enterprise Workflows: Better at handling ambiguous problems, staying productive over longer sessions, and managing complex multi-step tasks
Autonomy: More capable of working independently on challenging problems without constant supervision

Important Considerations

1. Less Capable Than Mythos While Opus 4.7 represents an advancement, it remains less capable than Claude Mythos Preview, Anthropic’s most powerful model. Mythos Preview continues to have limited release due to safety concerns outlined in Project Glasswing.

2. Reduced Cyber Capabilities Anthropic reduced the model’s cyber capabilities during training compared to Mythos Preview and implemented safeguards that automatically detect and block requests indicating prohibited or high-risk cybersecurity uses.

3. Token Usage Changes

Uses an updated tokenizer that can result in 1.0 to 1.35 times more tokens for the same input
May produce more output tokens at higher effort levels due to increased reasoning

4. New Effort Level Anthropic introduced a new “xhigh” effort level between high and max settings, providing additional control over the balance between reasoning capability and response speed.

Pricing and Availability

Pricing remains the same as Opus 4.6: $5 per million input tokens and $25 per million output tokens. The model is available across all Claude products, the Claude API, Amazon Bedrock, Google Cloud’s Vertex AI, and Microsoft Foundry.

Summary

Opus 4.7 is a significant upgrade that excels particularly in complex coding tasks, vision capabilities, and professional work quality. The feedback from early adopters has been very positive, with many calling it a “clear step up” and noting meaningful improvements in real-world applications. However, it’s positioned as a safer, more broadly available model compared to the more powerful but restricted Mythos Preview.

References:

Back Donate