Two professional-level certifications now target the developer who puts generative AI into production. AWS Certified Generative AI Developer - Professional (AIP-C01) validates integrating foundation models into applications with AWS services. Anthropic's Claude Certified Developer - Professional, listed as a preview, validates running Claude-based systems in production at scale. Both cover integration, agents, evaluation, safety, monitoring and cost, so the question is less which one is harder and more which platform your work runs on. This comparison sets the two side by side using each vendor's own published details.
Status on November 16, 2026
AIP-C01 is live, with a published exam guide. Claude Certified Developer - Professional is a preview entry in Anthropic's Partner Academy catalog: format, price and eight weighted domains are listed, but there is no exam guide, registration or launch date yet. "CCDV-P" is Preporato's label; Anthropic has not published a code. We will update this page when Anthropic publishes the guide.
The two exams at a glance
Claude Developer - Professional and AWS AIP-C01 side by side
| Claude Certified Developer - Professional | AWS Certified Generative AI Developer - Professional (AIP-C01) | |
|---|---|---|
| Vendor | Anthropic | Amazon Web Services |
| Status | Preview: no registration yet | Live |
| Questions | 70 (60 scored, 10 unscored) | 75 (65 scored, 10 unscored) |
| Time | 135 minutes | 180 minutes |
| Price | $175 | $300 |
| Passing score | 720 (scaled 100 to 1,000) | 750 (scaled 100 to 1,000) |
| Question types | Multiple choice and multiple response | Multiple choice and multiple response |
| Languages | English (as listed) | English, Japanese, Korean, Simplified Chinese |
| Delivery | Pearson VUE, test center or online (as for the live Claude exams) | Pearson VUE test center or online proctored |
| Validity | 12 months | 3 years |
| Who can register | Live Claude exams are open to people at Claude Partner Network organizations | Anyone |
| Domains | 8, from 20% down to 8% | 5, from 31% down to 11% |
| Platform | Claude Platform, Agent SDK and Claude Code | AWS services, mainly Amazon Bedrock, across many foundation models |
Three differences matter more than the rest.
Platform scope. AIP-C01 is about AWS: its skills name Amazon Bedrock, SageMaker AI, Step Functions, Lambda, OpenSearch Service, CloudWatch and dozens of other services, and the models are whichever foundation models you run on them. Claude Developer - Professional is about one model family and its tooling in depth: the Claude API, the Agent SDK and Claude Code.
Access. Anyone can register for AIP-C01 today. The live Claude exams are open to people at Claude Partner Network organizations, and Developer - Professional will very likely work the same way.
Renewal. An AWS certification lasts three years, and AWS lists passing the latest exam as the way to renew AIP-C01. The live Claude certifications last 12 months, with a free renewal assessment before they expire; whether Developer - Professional will offer one is not stated yet.
Try 15 CCDV-P practice questions free
They come from the CCDV-P practice tests, and each option shows its explanation once you answer.
- 6 timed practice exams
- Every answer explained
- A score for each domain
- $19.99 one-time
What each one certifies
AWS says AIP-C01 "validates a candidate's ability to effectively integrate foundation models (FMs) into applications and business workflows" and demonstrates "practical knowledge of how to implement GenAI solutions into production environments by using AWS technologies." Its target candidate has two or more years of building production-grade applications on AWS or with open-source technologies, general AI/ML or data engineering experience, and a year of hands-on generative AI work. AWS lists model development and training, advanced ML techniques, and data and feature engineering as out of scope.
Anthropic says Developer - Professional "validates running Claude-based systems in production at scale: instrumenting them with observability and monitoring, diagnosing complex failures and leading incident response, gating releases with evaluation, designing resilient integrations, controlling cost and latency across a fleet, hardening tool-using agents with layered guardrails, orchestrating durable multi-agent systems, and operating fleets of coding agents for software delivery."
Read together: AIP-C01 leans toward designing and building generative AI solutions on AWS, with retrieval and data at its center. Developer - Professional leans toward operating Claude systems after launch, with observability, evaluation and incident response at its center.
The domains side by side
Domain weights
| Claude Developer - Professional | Weight | AWS AIP-C01 | Weight |
|---|---|---|---|
| Production Integration, Reliability, and Knowledge Grounding | 20% | Foundation Model Integration, Data Management, and Compliance | 31% |
| Advanced Agent Architecture and Orchestration | 17% | Implementation and Integration | 26% |
| Observability and Production Monitoring | 12% | AI Safety, Security, and Governance | 20% |
| Evaluation Engineering and Quality Assurance | 12% | Operational Efficiency and Optimization for GenAI Applications | 12% |
| Runtime Guardrails and Production Security | 11% | Testing, Validation, and Troubleshooting | 11% |
| Advanced Debugging and Incident Response | 10% | ||
| Agent Fleet Operations for Software Delivery | 10% | ||
| Cost and Latency Optimization at Scale | 8% |
The domain names do not line up one to one, so the table below matches each topic to where it appears in each exam. For AIP-C01, the task numbers come from the AWS exam guide. For Developer - Professional, the domain is as listed; Anthropic's task statements arrive with its exam guide.
Where each topic lives
| Topic | Claude Developer - Professional | AWS AIP-C01 |
|---|---|---|
| Retrieval and grounding | Part of Domain 1 (20%), next to integration reliability | Tasks 1.4 and 1.5: vector stores and retrieval, a large share of the 31% domain |
| Resilient API integration | Domain 1: errors, retries, rate limits, context windows | Task 2.4: exponential backoff, rate limiting, fallbacks, streaming |
| Agents and orchestration | Domain 2 (17%): subagents, compaction, durable multi-agent systems | Task 2.1: agentic AI with Strands Agents, Agent Squad, Step Functions and MCP |
| Monitoring | Domain 3 (12%): usage data, request IDs, telemetry | Task 4.3: CloudWatch, model invocation logs, tracing |
| Evaluation | Domain 4 (12%): evals as release gates | Task 5.1: evaluation systems, including LLM-as-a-judge and quality gates |
| Safety and security | Domain 5 (11%): layered guardrails for tool-using agents | Domain 3 (20%): safety controls, data privacy, governance, responsible AI |
| Troubleshooting | Domain 6 (10%): debugging and incident response | Task 5.2: context window overflow, integration, prompt and retrieval issues |
| Coding agents in delivery | Domain 7 (10%): fleets of coding agents in CI | No domain; developer productivity with Amazon Q Developer appears in Task 2.5 |
| Cost and latency | Domain 8 (8%): caching, batching, routing at fleet scale | Tasks 4.1 and 4.2: token efficiency, model selection, caching, batching |
| Model deployment and customization | Not listed | Tasks 1.2 and 2.2: model selection, fine-tuned model deployment, provisioned throughput |
Where they overlap
The overlap is real, and it is the transferable core of production generative AI work:
- Resilience. Both expect retries with exponential backoff, handling of rate limits, streaming for long responses, and a plan for when the model or service pushes back.
- Evaluation. Both treat evaluation as an engineering system: automated quality gates before deployment, regression tests on model outputs, and model-graded evaluation where code checks cannot judge.
- Monitoring and troubleshooting. Both want token usage, latency and quality tracked per request, and both include context-window overflow among the failures you diagnose.
- Cost. Both cover token efficiency, prompt caching, batching and routing simpler work to cheaper models.
- Agents and tools. Both include multi-agent systems and the Model Context Protocol.
If you prepare for one, these parts carry over. The concepts are shared; the services, parameters and error codes you are tested on are each vendor's own.
Where they differ
AIP-C01 goes wider. Its largest domain covers solution design, model selection, data pipelines, vector stores and retrieval, and prompt governance. Its skills span deployment choices that a single-vendor exam does not need, such as provisioned throughput, SageMaker AI endpoints, fine-tuned model lifecycles and cross-Region inference. Enterprise integration (API Gateway, EventBridge, CI/CD for GenAI components) is a task of its own.
Developer - Professional goes deeper on operations. Observability, evaluation, guardrails and debugging are four separate domains that add up to 45% of its weight, and incident response is named in the exam description. It also gives a whole domain to running coding agents unattended in delivery pipelines, which AIP-C01 has no equivalent for.
The study material differs in kind. AIP-C01 preparation is mostly AWS service knowledge: which service fits which requirement and how they connect. Developer - Professional preparation is Claude platform behavior: what a given error, stop reason, setting or usage number means, and what to do about it.
Which one should you take?
Pick by your situation
| Your situation | Start with |
|---|---|
| Your generative AI systems run on AWS and use several model providers through Bedrock | AIP-C01 |
| You build and operate products on the Claude API or the Agent SDK directly | Claude Developer - Professional (once it opens) |
| Your team runs Claude Code across many developers or in CI | Claude Developer - Professional |
| You run Claude models through Amazon Bedrock | Either; AIP-C01 for the AWS side, Developer - Professional for Claude itself |
| You need a certification now, without partner access | AIP-C01, which is open to anyone today |
| Your employer is a Claude Partner Network firm delivering Claude projects | Claude Developer - Professional, after CCDV-F if you hold neither |
Many teams run Claude on Amazon Bedrock, and for them the two certifications are complementary. AIP-C01 covers the AWS side (IAM boundaries, Step Functions orchestration, CloudWatch, Bedrock Guardrails and Knowledge Bases), while Developer - Professional covers Claude's own behavior and tooling, including Claude Code, which runs on Bedrock as well as on the Claude API.
Taking both
If you plan to hold both, the order depends on access and timing more than difficulty. AIP-C01 is open now, so engineers without partner access often start there. Developer - Professional has no launch date, and its preparation rewards hands-on Claude work, so the months before it opens are a good time to build the runbooks, eval gates and CI agents its domains describe.
Study the shared core once: resilience patterns, evaluation design, monitoring and cost control. Then spend the remaining time on each vendor's specifics, AWS services for AIP-C01 and Claude Platform and Claude Code behavior for Developer - Professional.
How to prepare for each
For AIP-C01: read the AWS exam guide task by task, then the AIP-C01 complete guide and the Amazon Bedrock guide for AIP-C01. Preporato's AIP-C01 practice exams give you timed full-length practice.
For Claude Developer - Professional: start with the exam guide and the 20 free practice questions, follow the 6-week study plan, and keep the cheat sheet at hand. Preporato's Developer - Professional practice exams are six Early Access forms of 70 questions, written to Anthropic's listed weights and revised to the guide when it is published.
Frequently asked questions
The short version
AIP-C01 certifies building generative AI on AWS across many models, with retrieval, data and AWS services at its core, and you can take it today. Claude Certified Developer - Professional will certify operating Claude systems in production, with observability, evaluation, guardrails, incident response and coding agents at its core, once Anthropic opens it. If your systems run on AWS, start with AIP-C01; if they run on Claude, prepare for Developer - Professional; if they run Claude on AWS, the two complement each other.
Sources: