Skip to main content
Back to challenges
Challenge

High-Performance Reasoning Swarm with Mastra AI

Construct a distributed agent system designed to handle complex media analysis tasks. This challenge focuses on high-speed inference and orchestration using Groq Cloud and Modal to scale reasoning tasks. The system utilizes GPT-5.4 Pro for synthesis and Claude Sonnet 4.6.6 for creative content formatting, coordinated entirely within the Mastra AI framework to ensure consistent execution.

Special Purpose AgentsHosted by Vera
Sign in free to start
Challenge brief

What you are building

The core problem, expected build, and operating context for this challenge.

Orchestrate a high-performance multi-model agent swarm using Mastra AI to analyze and synthesize media content

Delivery guide

How work is evaluated

Evaluation

Evaluation of swarm throughput and reasoning accuracy

Datasets

Challenge datasets

Public datasets, plus the private datasets you have access to.

Evaluation rubric

How submissions are scored

These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.

Dimensions
2 scoring checks
Binary
2 pass or fail dimensions
Ordinal
0 scaled dimensions
Dimension 1

Model Switch

Verify model fallback

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Dimension 2

Inference Latency

Average seconds per task • target: 2 • range: 0-10

Binary check

This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.

Learning goals

What you should walk away with

  • Master Mastra AI agent orchestration to manage agent state and workflow transitions

  • Deploy serverless inference containers on Modal to handle specialized high-compute sub-tasks

  • Implement model routing to use Groq Cloud for fast-path processing versus GPT-5.4 Pro for deep reasoning

  • Build a Claude Sonnet 4.6.6 synthesis node for generating high-quality media briefing reports

  • Create custom tools within Mastra AI to interface with media APIs for real-time analysis

  • Implement observability hooks to monitor token usage and reasoning depth across the swarm

Resources and assets

Reference links and supporting material

Dataset notes

Sample data for 1 tasks

Challenge setup

Evaluation of swarm throughput and reasoning accuracy

Challenge input

Media transcript

Mastra AI

TypeScript agent framework.

GPT-5

Policy Serving in the agent workflow.

Groq Cloud

Low-latency inference cloud.

Evaluated output

Synthesized report

Evaluation checks
  • Verify model fallback
  • Average seconds per task • target: 2 • range: 0-10
Configured evaluation evidence
  • Inference Latency target: 2
  • 1 public reference case
Configured runtime
  • TypeScript execution harness

Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.

View technical recipe

Configured tools

Action Space
  • Mastra AI · Required
Policy Serving
  • GPT-5 · Optional
  • Groq Cloud · Optional
Orchestration
  • Mastra AI · Required

Evaluation contract

  • Model Switch · Weight 1
  • Inference Latency · Weight 1

Metrics

  • Inference Latency · Target 2

Reference cases

  • 1 public reference case

Workflow tasks

  1. Task 1

    Input: Media transcript

    Output: Synthesized report

    Execution harness: TypeScript

Recipe freeze marker

No recipe freeze time is recorded.

Frequently Asked Questions about High-Performance Reasoning Swarm with Mastra AI