---
type: "post"
id: "anthropic-s-claude-mythos-5-2-is-an-07a4f8"
slug: "anthropic-s-claude-mythos-5-2-is-an-07a4f8"
title: "Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026"
description: "Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026, but access is limited to organizations verified through Anthropic programs."
retrieval_nugget: "Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026, but access is limited to organizations verified through Anthropic programs. Its documentation lists a 1-million-token context window, 128,000-token maximum output, and pricing of $8 per million input tokens and $40 per million output tokens. The combination matters for teams assessing access eligibility and long-context costs."
published_at: "2026-11-11"
updated_at: "2026-11-11"
record_date: "2026-11-11"
date_kind: "published_at"
topics: ["claude","anthropic","google"]
entities: ["Anthropic","Claude Mythos 5.2","Claude Fable 5.2","Claude API","Amazon Bedrock","Google Cloud","Microsoft Foundry","Cyber Verification Program"]
related_patterns: ["agent-economics-moves-to-completed-work"]
novelty_label: "new_fact"
what_changed: "Claude Mythos 5.2’s specifications and pricing match Fable 5.2, including $8/$40 per million input/output tokens, and access is limited to organizations verified through Anthropic programs."
prior_refs: []
source_url: "https://platform.claude.com/docs/en/models/mythos-5-2/overview"
source_urls: ["https://platform.claude.com/docs/en/models/mythos-5-2/overview"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:anthropic-s-claude-mythos-5-2-is-an-07a4f8"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/anthropic-s-claude-mythos-5-2-is-an-07a4f8/","markdown":"https://newruntime.com/posts/anthropic-s-claude-mythos-5-2-is-an-07a4f8.md","json":"https://newruntime.com/posts/anthropic-s-claude-mythos-5-2-is-an-07a4f8.json"}
markdown_url: "https://newruntime.com/posts/anthropic-s-claude-mythos-5-2-is-an-07a4f8.md"
---

# Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026

## Retrieval answer

Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026, but access is limited to organizations verified through Anthropic programs. Its documentation lists a 1-million-token context window, 128,000-token maximum output, and pricing of $8 per million input tokens and $40 per million output tokens. The combination matters for teams assessing access eligibility and long-context costs.

## Summary
Anthropic’s Claude Mythos 5.2 is an active model released on November 9, 2026, but access is limited to organizations verified through Anthropic programs. Its documentation lists a 1-million-token context window, 128,000-token maximum output, and pricing of $8 per million input tokens and $40 per million output tokens. The combination matters for teams assessing access eligibility and long-context costs.

## What happened
Anthropic’s Claude Platform documentation lists Claude Mythos 5.2 as released on November 9, 2026, with active status and verification required. The page says organizations must be verified through one of Anthropic’s verification programs, such as the Cyber Verification Program, and directs applicants to apply to the program covering their use case. This is not open general availability: eligibility is a condition of access.

The listed model IDs are `claude-mythos-5-2` for the Claude API, Amazon Bedrock, Google Cloud in both Google-managed and Anthropic-managed configurations, and Microsoft Foundry. The documentation sets a retirement date no sooner than November 9, 2027. It also describes Mythos 5.2 as having the same specifications and pricing as Claude Fable 5.2.

The specification table gives Mythos 5.2 a 1-million-token context window and a maximum output of 128,000 tokens. Its input is text and images and its output is text. Thinking is adaptive and always on, with default effort set to `high`; comparative latency is listed as moderate. Both the reliable knowledge cutoff and training-data cutoff are June 2026.

Pricing is $8 per million input tokens and $40 per million output tokens. Prompt-cache writes cost $10 per million tokens for a five-minute cache and $16 per million for a one-hour cache; cache reads cost $0.20 per million. Batch API input and output receive a 50% discount. The page points to a fast mode for Messages API requests at premium pricing, but does not state the premium rate here. It also links to a system card and a migration guide from Mythos 5.1.

## What the archive adds
New Runtime covered Claude Fable 5.2’s formalization of Fermat’s Last Theorem in an earlier note dated November 9, 2026: [Introducing Claude Fable 5.2 and Claude Mythos 5.2](https://newruntime.com/posts/introducing-claude-fable-5-2-and-claude-mythos-6b1c41/). This documentation adds model-specific operational details, including Mythos access requirements, platform IDs, pricing, and limits; it is a new fact on a known release story.

## Trend signal
This supports the pattern that [agent economics moves to completed work](https://newruntime.com/patterns/agent-economics-moves-to-completed-work/) only weakly: the documentation makes token costs, batch discounts, and cache rates explicit, but reports no completed-task cost or outcome measure. The immediate signal is that builders can compare usage modes, not that the economics have shifted to verified task completion.

## Why now
The documentation dates the release to November 9, 2026, and states a retirement commitment of no sooner than November 9, 2027. The release timing gives teams a concrete point to evaluate whether this restricted model is eligible for their use case and worth testing against existing options.

## Connections to other events
New Runtime’s earlier coverage dated November 9, 2026 reported that Claude Fable 5.2 formalized Fermat’s Last Theorem in 30 hours. The present model documentation connects that release context to Mythos 5.2’s access conditions and API economics; it does not establish that the formalization result was produced by Mythos specifically.

## Why it matters
A million-token context window and 128,000-token output ceiling can accommodate large inputs and substantial responses, but neither number alone establishes quality, latency in a particular workload, or cost-effectiveness. At the documented rates, teams should model both input and output usage, along with cache writes and reads, before committing a workload. Batch processing can halve input and output token charges where its execution model fits.

That cost model makes workload shape important. A team passing long, frequently reused context may need to compare the cost of cache creation with the much lower cache-read rate, rather than treating the headline input price as the entire bill. A workload that can tolerate Batch API processing has a separate pricing path from an interactive Messages API workflow. The documentation does not say which workload types will produce the best practical results, so these are design variables to test rather than deployment guarantees.

Access is the first practical constraint. Organizations cannot assume they can call the model simply because an API identifier and several hosting platforms are listed; Anthropic says verification through an applicable program is required. Teams should confirm eligibility and test the model on representative tasks before designing a production dependency around it. The published lifecycle commitment also gives a minimum horizon, not a guarantee that the model will remain unchanged.

## Limits and uncertainty
Anthropic’s documentation provides specifications, listed prices, platforms, and access conditions, but does not quantify benchmark performance, task success, or actual latency beyond the label “moderate.” It does not state the fast-mode premium in this overview, nor does it establish how quickly an application to a verification program will be processed. The system card is linked, but the model page alone does not independently validate safety or performance claims. The listed retirement date is a minimum commitment, not a promise about future revisions or aliases.

## Reader takeaway
Mythos 5.2 is a restricted-access model with a large context window and explicit usage rates, not a generally available upgrade teams can assume they can deploy. Check program eligibility first, then compare its cost and behavior with your current model on a representative workload. The matching Fable 5.2 pricing makes careful model selection and workload testing especially important.

## Open question
How selective are Anthropic’s verification programs, and what use cases will they approve for Mythos 5.2?

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Claude](https://newruntime.com/topics/claude.md): Explore the claude topic hub.
- topic hub: [Anthropic](https://newruntime.com/topics/anthropic.md): Explore the anthropic topic hub.
- topic hub: [Google](https://newruntime.com/topics/google.md): Explore the google topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Sources](https://newruntime.com/sources/): See the original sources this archive cites.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)