Claude Mythos 5

Claude Mythos 5 is the successor to Claude Mythos Preview and is available only to approved customers in Project Glasswing, not for general self-serve use.

Platform: Replicate

Reasoning AIRestricted AccessCoding AgentLong-horizon Workflows

0 runs

Input: $10 per million tokens; Output: $50 per million tokens

Commercial

🚀Function Overview

A very high-capability Anthropic model in a restricted access program, intended for controlled high-value workflows.

Key Features

Successor to Claude Mythos Preview, available via Project Glasswing
1M context and 128k max output values
Adaptive thinking is always on; extended thinking budgets are not used
Comparable API surface to Claude Fable 5 with stricter access control
Intended for restricted, invitation-only usage rather than broad general availability

Use Cases

•Defensive-critical infrastructure use under controlled governance
•Research or evaluation environments requiring frontier-level model capacity
•Advanced long-horizon planning and analysis workflows in approved accounts
•High-precision coding tasks where risk controls or governance context are pre-defined

⚙️Input Parameters

prompt

string

The prompt to send to the model

system_prompt

string

System prompt for policy and workflow constraints

max_tokens

integer

Maximum number of output tokens to generate

effort

string

Reasoning effort for adaptive thinking controls

max_image_resolution

number

Optional image-resolution cap (megapixels) to reduce image token consumption

💡Usage Examples

Example 1

Input Parameters

{
  "prompt": "Create a secure patching plan for a critical third-party dependency and rank remediation priorities under incident constraints.",
  "system_prompt": "Assume a strict governance context and propose only defensible, auditable actions.",
  "effort": "high",
  "max_tokens": 4096,
  "max_image_resolution": 0.5
}

Output Results

I would start by mapping the dependency graph for transitive risk concentration and operational blast radius.

Then I would separate vulnerabilities by exploitability, deployment criticality, and rollback impact.

Priority 1: patch the package in the release-critical path with minimal semantic drift risk.

Priority 2: add temporary mitigations for non-blocking but exposed attack classes.

Priority 3: schedule follow-up dependency upgrades with compatibility validation and rollout safeguards.

Use a governance checkpoint after each step to avoid introducing avoidable operational debt.

Quick Actions

Use NowView Documentation

Technical Specifications

Hardware Type
Run Count: 0
Commercial Use: Supported
Pricing: Input: $10 per million tokens; Output: $50 per million tokens
Platform: Replicate

Related Keywords

Project GlasswingInvitation-only modelRestricted Access AILong-horizon ReasoningAnthropic frontier modelHigh-autonomy coding

Related Models

Bielik 1.5B v3 Instruct

Bielik-1.5B-v3-Instruct is a generative text model featuring 1.6 billion parameters. It is result of collaboration between the open-science/open-souce project SpeakLeash and the High Performance Computing (HPC)

Cordia-A6 Text Generation Model

A model for generating text sequences based on input prompts and adjustable parameters.

Claude Sonnet 4

Claude Sonnet 4 is a significant upgrade to 3.7, delivering superior coding and reasoning while responding more precisely to your instructions