What the family record covers
Anthropic documents Claude as a model family with separate current models for different workload profiles. Current models accept text and image input and return text, but their identifiers, reasoning behavior, cloud availability, and lifecycle rules must be checked independently.
Do not infer the model behind a Claude product session from the product name alone. For evaluation, record the API model identifier or product surface and separate vendor descriptions from observed task results.
Documented access paths
- Claude API
- Documented cloud-platform distribution, subject to model and region availability
- Distinct model identifiers for different capability and latency profiles
Capabilities to verify
- Long-document instruction adherence and evidence traceability
- Coding or agentic task completion under bounded permissions
- Image interpretation on domain-specific visual material
- Behavior consistency across the exact first-party or cloud endpoint in scope
Decision questions
- Which Claude model ID and hosting surface will be approved?
- Does the workload benefit from a capability-oriented model or a faster operating profile?
- How will the team retest before a model migration or deprecation date?
Evaluation cautions
- Claude product access and Claude API model access are different decision surfaces.
- Cloud identifiers and routing choices can differ from first-party API identifiers.
- Published context or capability limits do not prove reliability across a full document set.
Related decisions
Move from a family name to the product, provider, workflow, comparison, or protocol that defines the real choice.
Sources checked
Open the original pages before relying on a time-sensitive product decision.