--- title: Claude Fable 5 on AWS (June 2026): Mythos-Class Models, Safeguards, and What Changes for Bedrock Teams description: On June 9, 2026 AWS GA’d Claude Fable 5 on Bedrock — 1M-token context, 128K output, long-running autonomous work, and a data-retention opt-in that blocks some regulated workloads. Here is the fit decision, not the press release. url: https://www.factualminds.com/blog/claude-fable-5-aws-bedrock-mythos-class-guide-2026/ datePublished: 2026-06-10T00:00:00.000Z dateModified: 2026-06-10T00:00:00.000Z author: Palaniappan P category: Generative AI tags: bedrock, claude, anthropic, foundation-models, genai, aws, agentic-ai --- # Claude Fable 5 on AWS (June 2026): Mythos-Class Models, Safeguards, and What Changes for Bedrock Teams > On June 9, 2026 AWS GA’d Claude Fable 5 on Bedrock — 1M-token context, 128K output, long-running autonomous work, and a data-retention opt-in that blocks some regulated workloads. Here is the fit decision, not the press release. On **June 9, 2026**, AWS [announced](https://aws.amazon.com/about-aws/whats-new/2026/06/claude-fable-5-aws/) general availability of **Claude Fable 5** on [Amazon Bedrock](https://aws.amazon.com/bedrock/) and [Claude Platform on AWS](https://aws.amazon.com/claude-platform/) — Anthropic's first **Mythos-class** model available to all customers, with safeguards designed for broader enterprise use. The [AWS News Blog post](https://aws.amazon.com/blogs/aws/anthropic-claude-fable-5-on-aws-mythos-class-capabilities-with-built-in-safeguards-now-available/) frames it as a step-change in long-running knowledge work, software engineering, and vision — not another incremental Sonnet refresh. If you run production GenAI on Bedrock today, the announcement changes three things that a feature bullet list will not surface: **(1)** you must opt into provider data sharing before the first invoke, **(2)** refusal handling becomes a primary code path, and **(3)** the unit of work shifts from "one API call" to "one autonomous job that may run for hours." This post is the field guide — what Fable 5 actually is, how it differs from Mythos 5 and Claude 4.6, and which workloads should move first. > **Benchmark pattern (not a cited client)** — Modeled a compliance-heavy B2B SaaS: ~180 multi-page PDF contracts/month through a Bedrock extraction pipeline, previously **Claude 3.5 Sonnet** at ~**$3,900/mo** inference (mix of on-demand + partial prompt cache). A two-week Fable 5 pilot on the vision-heavy path (nested tables in exhibits) cut human rework from **~11 hrs/week → ~3 hrs/week** on a 12-document sample, at roughly **2.1×** per-document inference cost vs Sonnet on the same corpus. Net: positive ROI on the document lane only; the chat/classification lane stayed on Nova Lite. --- ## What "Mythos-class" means in practice Prior Claude generations were optimized for strong single-turn and short multi-turn reasoning. **Mythos-class** (Anthropic's term for the capability tier Fable 5 brings to GA) targets a different failure mode: models that lose the thread on **long-horizon** work — multi-file refactors, research synthesis across dozens of sources, or document pipelines where the model must plan, execute, self-check, and continue without a human re-prompting every ten minutes. Per the [Bedrock model card](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-fable-5.html), Fable 5 ships with: | Spec | Value | Why it matters architecturally | | ---------------- | --------------------------------- | --------------------------------------------------------------------------------------------- | | Context window | **1M tokens** | Whole codebases + doc corpora in one session — but you still pay for what you send | | Max output | **128K tokens** | Long generated artifacts (specs, migration plans) without chunking — watch streaming timeouts | | Knowledge cutoff | **January 2026** | Newer than Sonnet 4.6 family for recent AWS/Anthropic features | | Reasoning | **Adaptive thinking (always on)** | Cannot disable; adds latency vs "fast mode" Sonnet calls | | Input modalities | Text + **image** | Diagrams, charts, nested PDF tables — the vision story AWS highlights for finance/legal | | Launch date | **June 9, 2026** | Standard inference tier; Priority/Flex/Reserved not listed at launch | The capabilities AWS emphasizes in the announcement — **long-running asynchronous execution**, **advanced vision**, and **proactive self-verification** (updating skills, building evaluation harnesses, checking its own output) — are not marketing adjectives. They imply your orchestration layer must support **checkpointing, idempotent retries, and human-in-the-loop gates** on jobs that outlive a single Lambda timeout or API Gateway limit. --- ## Fable 5 vs Claude Mythos 5: the safeguard split AWS and Anthropic ship two names for nearly the same frontier capability: - **Claude Fable 5** — GA on Bedrock. Includes **safeguards**: in sensitive domains (cybersecurity, biology, chemistry, health), harmful prompts may receive a response from a **less capable model (Opus 4.8)** instead of full Fable capability. Blocking classifiers also return `stop_reason: "refusal"` for dual-use content. - **Claude Mythos 5** — Same model **without** those limits. Available only to a **small group of vetted customers** who already had Mythos Preview access. **We recommend Fable 5** for any workload that touches production customer data or faces external auditors. Mythos 5 is a research and red-team surface, not a default production route. If your security team asks "can we get the unrestricted model," the honest answer is: not through normal enterprise procurement — and you probably should not want it on a customer-facing path. --- ## Two front doors on AWS: Bedrock vs Claude Platform on AWS | Concern | Amazon Bedrock | Claude Platform on AWS | | -------------- | ---------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ | | Governance | IAM, SCPs, CloudTrail, PrivateLink | AWS billing + auth; Anthropic-native UX | | Composition | Guardrails, Knowledge Bases, Agents, Flows | Anthropic platform features first | | API surface | Converse, Invoke, Messages API on `bedrock-runtime` / `bedrock-mantle` | Claude Platform APIs (see [docs](https://docs.aws.amazon.com/claude-platform/latest/userguide/welcome.html)) | | Data residency | Regional inference profiles (in-region / geo / global) | Confirm per Claude Platform on AWS guidance | **Pick Bedrock** when the app already lives inside your AWS compliance boundary. **Pick Claude Platform on AWS** when the buyer is a team that wants Anthropic's native experience but needs AWS procurement — not when you need to compose with an existing Bedrock RAG stack without re-plumbing. For invoke patterns, AWS documents both the **Anthropic Messages API** on `bedrock-mantle` and the **Converse API** on `bedrock-runtime`: ```python # Converse API — boto3, us-east-1, June 2026 model ID import boto3 bedrock = boto3.client("bedrock-runtime", region_name="us-east-1") response = bedrock.converse( modelId="global.anthropic.claude-fable-5", messages=[{"role": "user", "content": [{"text": "Your prompt"}]}], inferenceConfig={"maxTokens": 4096}, ) print(response["output"]["message"]["content"][0]["text"]) ``` Model IDs and regional endpoints are in the [model card](https://docs.aws.amazon.com/bedrock/latest/userguide/model-card-anthropic-claude-fable-5.html) — pin `anthropic.claude-fable-5` for strict in-region, or `global.anthropic.claude-fable-5` when residency allows global routing. --- ## The gate most teams miss: data retention opt-in You cannot invoke Fable 5 until you opt in via the **Data Retention API** — **no console UI at launch**. Set `provider_data_share` on the `bedrock-mantle` endpoint: ```bash curl -X PUT "https://bedrock-mantle.us-east-1.api.aws/v1/data_retention" \ -H "x-api-key: " \ -H "Content-Type: application/json" \ -d '{"mode": "provider_data_share"}' ``` Anthropic requires **30-day retention of inputs and outputs** and **human review** under this mode. That is the right trade for frontier safety at scale — and a **hard stop** for some regulated workloads until legal signs off. We have seen security reviews stall here while engineering already merged the model ID into staging. > **Reproduce this** — Copy [`data-retention-opt-in.sh`](https://www.factualminds.com/examples/architecture-blog-2026/claude-fable-5-bedrock/data-retention-opt-in.sh) and [`refusal-handler-example.py`](https://www.factualminds.com/examples/architecture-blog-2026/claude-fable-5-bedrock/refusal-handler-example.py) from `examples/architecture-blog-2026/claude-fable-5-bedrock/`. Run the opt-in script in a sandbox account after security review; run the Python handler to see refusal fallback wiring. If your org already runs [HIPAA-eligible Bedrock](/blog/hipaa-compliant-ai-aws-bedrock/) workloads, treat Fable 5 as a **new sub-process** — the BAA covers Bedrock; the **provider data share** is an additional disclosure decision, not a checkbox. --- ## Refusals are not errors — redesign your response path The model card is explicit: Fable 5's refusal rates are **materially higher** than previous Claude models. Classifiers return HTTP **200** with `stop_reason: "refusal"` and a `stop_details` category — not `4xx`. | Refusal timing | Billing | What your app should do | | --------------------------------------- | ------------------------------ | --------------------------------------------------------------------- | | Prompt-stage (blocked before inference) | Not billed | Log, route to human review or approved fallback model | | Mid-stream (partial output then block) | Billed for tokens before block | Persist partial output if safe; do not blindly retry identical prompt | > **What broke** — In the modeled pilot, **4 of 120** contract prompts (security-addendum language touching vulnerability disclosure) returned `refusal` on Fable 5 where Sonnet had completed normally. Downstream JSON schema validation failed because the handler treated 200 + refusal as success. Fix: branch on `stopReason` before parsing; queue refusals to a paralegal review bucket instead of retry loops. Retry loops would have doubled cost on mid-stream refusals. We recommend **Bedrock Guardrails** ([production setup guide](/blog/how-to-set-up-amazon-bedrock-guardrails-production/)) _in addition to_ Fable's built-in classifiers — defense in depth for PII and prompt injection, not a replacement for refusal handling. --- ## When to move workloads — and when to stay on Sonnet or Nova Use the decision matrix in [`model-selection-decision-matrix.md`](https://www.factualminds.com/examples/architecture-blog-2026/claude-fable-5-bedrock/model-selection-decision-matrix.md). Short version: **Move first to Fable 5:** - Long-horizon **coding agents** (multi-file refactors, test harness generation) where [Claude Code-style skills](/blog/terraform-claude-skill-aws-production-guide/) already proved value but hit context ceilings - **Vision-heavy document** pipelines (financial statements, legal exhibits, architecture diagrams in PDFs) - **Async research** jobs orchestrated through [Bedrock Agents](/blog/how-to-build-amazon-bedrock-agent-tool-use-2026/) or Step Functions with checkpointing **Stay on Claude 4.6 Sonnet / Haiku or [Nova](/blog/aws-bedrock-nova-models-guide/):** - Customer chat under **500 ms p99** latency pressure - High-volume classification and moderation - Workloads where **data-retention opt-in** is not approved - Cost-sensitive token volumes — Fable 5 is frontier-priced; [token budget discipline](/blog/aws-bedrock-cost-optimization-token-budgets-model-selection/) still applies **Opinionated take:** Fable 5 is not a Sonnet upgrade. It is a **job runner**. If your architecture diagram still shows "API Gateway → Lambda → Bedrock" as a single synchronous box, Fable 5 belongs behind a **queue and worker** (SQS + ECS/Lambda with extended timeout, or [AgentCore](/blog/amazon-bedrock-agentcore-production/) for managed agent runtime) — not in the hot request path. --- ## Composing with the Bedrock stack you already have Fable 5 inherits the Bedrock features you use for other Anthropic models — with launch caveats on service tiers: - **[Knowledge Bases](/blog/how-to-build-rag-pipeline-amazon-bedrock-knowledge-bases/)** — sensible for long document grounding; pair with prompt caching (1,024-token minimum checkpoint, up to 4 checkpoints, 5 min / 1 hr TTL per model card) - **Guardrails** — still your app-level PII/topic policy; Fable adds model-level classifiers underneath - **Agents and Flows** — natural fit for multi-step jobs; ensure tool-use loops have spend caps - **Cross-region inference** — use geo/global profiles for throughput; re-check [data residency](/blog/aws-data-residency-sovereignty-guide-2026/) if prompts contain regulated data With [OpenAI models now on Bedrock](/blog/amazon-bedrock-openai-models-codex-managed-agents/) (April 2026 preview), Fable 5 completes a "**route by job shape**" portfolio: OpenAI for Codex-style coding surfaces, Fable for long autonomous Anthropic work, Sonnet/Nova for volume. --- ## What to do this week 1. **Legal/security review** of `provider_data_share` — before any staging invoke. 2. **Enable opt-in** in a sandbox account using the Data Retention API (script in artifacts folder). 3. **Pilot one long-horizon job** — not your highest-QPS API. Measure refusal rate, cost per completed job, and human rework hours saved. 4. **Add `stop_reason: refusal` handling** to your Bedrock client wrapper — treat it like a first-class outcome. 5. **Pin model IDs** per environment (`anthropic.claude-fable-5` vs `global.anthropic.claude-fable-5`) and document the residency implication. 6. **Leave chat/classification on Sonnet or Nova** until the pilot proves ROI on the document or agent lane. **If you only do one thing:** Run the data-retention and refusal-handling review **before** you change the model ID in production config. The model works on day one; the governance and client-code paths are what break. --- ## What this post does not cover - **On-demand pricing per million tokens** — verify live rates on [Bedrock pricing](https://aws.amazon.com/bedrock/pricing/); frontier launches move faster than blog publish cycles. - **Claude Mythos 5 access** — vetted-customer only; no enterprise architecture should depend on it. - **Claude Platform on AWS** feature parity with anthropic.com — we focused on Bedrock because that is where most AWS-native apps compose. - **Benchmarks** — AWS cites Anthropic's internal evals; we did not independently reproduce SWE-bench or MMLU numbers. Our numbers are from the modeled pilot above, not a published third-party benchmark. **Related:** [Bedrock vs OpenAI API for enterprise](/blog/aws-bedrock-vs-openai-api-enterprise/) · [Bedrock cost optimization](/blog/aws-bedrock-cost-optimization-token-budgets-model-selection/) · [Building Bedrock Agents with tool use](/blog/how-to-build-amazon-bedrock-agent-tool-use-2026/) · [Generative AI on AWS consulting](/services/generative-ai-on-aws/) ## Related reading - [How to Set Up Amazon Q for Business with SharePoint and S3](/blog/how-to-set-up-amazon-q-for-business-sharepoint-s3/) ## FAQ ### What is Claude Fable 5 and how is it different from Claude Mythos 5? Claude Fable 5 is Anthropic’s first generally available Mythos-class model on AWS (announced June 9, 2026). It is built for extended autonomous knowledge work and coding — multi-stage planning, sub-agent delegation, and self-verification — with a 1M-token context window and up to 128K output tokens. Claude Mythos 5 is the same underlying capability without Fable’s safety classifiers; AWS limits Mythos 5 to a small set of vetted customers. For most enterprises, Fable 5 is the production path. ### What do I have to do before I can invoke Claude Fable 5 on Bedrock? You must opt in to provider data sharing via the Bedrock Data Retention API (`provider_data_share` mode). At launch there is no console UI — only an API call against the bedrock-mantle endpoint. Anthropic requires 30-day retention of inputs and outputs plus human review. If your security or legal team has not approved that, do not route production traffic to Fable 5 yet. ### How do refusals work on Claude Fable 5? Fable 5 includes blocking classifiers for dual-use content (notably cybersecurity and biology). When a classifier fires, the API returns HTTP 200 with stop_reason "refusal" — not an error code. Refusal rates are materially higher than on prior Claude models. Prompt-stage refusals are not billed; mid-stream refusals bill tokens generated before the block. Production apps must handle refusal as a primary response path, with logging and a defined fallback or human-review route. ### Should we replace Claude 4.6 Sonnet with Fable 5 for all Bedrock workloads? No. Fable 5 is for long-running, document-heavy, or vision-intensive work where prior models lose coherence or need constant human steering. Short chat, classification, and high-QPS APIs should stay on Sonnet, Haiku-class models, or Nova for cost and latency. Fable 5 also requires the data-retention opt-in that Sonnet workloads may not need. Route by workload shape, not by "newest model." ### Bedrock or Claude Platform on AWS — which should we use? Use Amazon Bedrock when you need IAM, PrivateLink, regional data residency, and composition with Guardrails, Knowledge Bases, and Agents inside your existing AWS governance. Use Claude Platform on AWS when teams want Anthropic’s native platform experience (projects, Claude-style workflows) with unified AWS billing. Same model family; different control plane and developer UX. ### When should we NOT adopt Claude Fable 5 yet? Skip or delay if: (1) legal has not approved 30-day provider data retention and human review, (2) you have not measured refusal rates on your real prompt distribution, (3) the workload is a drop-in Sonnet replacement on short prompts, (4) you need Priority/Flex/Reserved throughput tiers (Fable 5 may be Standard-only at launch), or (5) prompts in security/bio/chem/health domains need predictable frontier capability — Fable may route those to a less capable fallback model. ### What regions support Claude Fable 5 on Bedrock? In-region inference is available in us-east-1 and eu-north-1 at launch. Geo and global inference profiles extend coverage across other US and EU regions — confirm the current matrix in the AWS model card before designing for strict in-region residency. Model IDs include anthropic.claude-fable-5 (in-region), us/eu.anthropic.claude-fable-5 (geo), and global.anthropic.claude-fable-5 (global). --- *Source: https://www.factualminds.com/blog/claude-fable-5-aws-bedrock-mythos-class-guide-2026/*