Caching fails for Fable 5 on Bedrock

[kblissett]

What happened?

pi relies on the following lines to determine if a model supports caching on Amazon Bedrock:

pi/packages/ai/src/providers/amazon-bedrock.ts

Lines 620 to 637 in 1da9039

	function supportsPromptCaching(model: Model<"bedrock-converse-stream">): boolean {
	const candidates = getModelMatchCandidates(model.id, model.name);
	const hasClaudeRef = candidates.some((s) => s.includes("claude"));
	if (!hasClaudeRef) {
	// Application inference profiles don't contain the model name in the ARN.
	// Allow users to force cache points via environment variable.
	if (typeof process !== "undefined" && process.env.AWS_BEDROCK_FORCE_CACHE === "1") return true;
	return false;
	}
	// Claude 4.x models (opus-4, sonnet-4, haiku-4)
	if (candidates.some((s) => s.includes("-4-"))) return true;
	// Claude 3.7 Sonnet
	if (candidates.some((s) => s.includes("claude-3-7-sonnet"))) return true;
	// Claude 3.5 Haiku
	if (candidates.some((s) => s.includes("claude-3-5-haiku"))) return true;
	return false;
	}

The model ID global.anthropic.claude-fable-5 does not match any of these patterns and thus no tokens are ever cached. This can lead to a dramatic increase in costs.

Steps to reproduce

Use global.anthropic.claude-fable-5 from provider amazon-bedrock.

Expected behavior

pi should use normal caching behavior when this model is selected.

Version

0.79.1

[github-actions]

This issue was auto-closed. All issues from new contributors are auto-closed by default.

Maintainers review auto-closed issues daily and reopen worthwhile ones. Issues that do not meet the quality bar in CONTRIBUTING.md will not be reopened or receive a reply.

If a maintainer replies lgtmi on one of your issues, your future issues will stay open. If a maintainer replies lgtm, your future issues and PRs will stay open.

See CONTRIBUTING.md.