What are the verified coding and reasoning benchmarks for GPT 5 Mini?

GPT 5 Mini achieves SWE-bench Verified: 54% (Thinking), LiveCodeBench: 58.5%, AIME: 76%. Verified sources: OpenAI GPT-5-mini Specifications (OpenAI).

How do I run GPT 5 Mini with Swarm?

Configure GPT 5 Mini via CLI using: swarm ctl model set --provider openai --model openai/gpt-5-mini, or switch models interactively inside the terminal TUI with /models.

Does Swarm isolate GPT 5 Mini code edits from other agents?

Yes. Swarm provisions isolated Git worktrees for every session and subagent, preventing file race conditions and keeping your main working tree clean until verification passes.

Can GPT 5 Mini be assigned to specific subagent roles?

Yes. In ~/.swarm/config.json, you can configure GPT 5 Mini as Primary/Planner (DAG task decomposition), Coder (implementation in isolated worktrees), or Finder (read-only symbol search).

GPT 5 Mini — Specs, Context & Capabilities

Q: What is the recommended Swarm multi-agent topology for GPT 5 Mini?

Swarm recommends GPT 5 Mini as Coder Subagent (Tier 3: High-Throughput Subagent Engine). Strong coding benchmarks allow running 4-8 parallel Coder workers across separate worktrees simultaneously. Best paired with: Claude 3.7 Sonnet or GPT-5 as Lead Planner.

GPT 5 Mini Benchmarks & Agentic Accuracy

Verified primary research & developer preference metrics

Primary Sources Verified

SWE-bench Verified

54%

ThinkingOpenAI Technical Report ↗

LiveCodeBench

58.5%

ThinkingLiveCodeBench ↗

LMSYS Coding Elo

1365Elo

Coding ArenaLMSYS Arena ↗

AIME 2024 / 2025

76%

ThinkingOpenAI Technical Report ↗

Pricing ($ / 1M tokens)

$0.4/ $1.6

$0.1 cached readStandard API Pricing

What This Means in Real Software Development

Exceptional 54.0% SWE-bench Verified score at 1/6th the cost of frontier flagship models.
Blazing throughput slashes execution times across multi-agent task programs.
Scores 76.0% on AIME 2024, handling intricate algorithmic edge cases reliably.
Ideal engine for parallel Coder subagents running simultaneous test suites.

Swarm Multi-Agent Topology

Tier 3: High-Throughput Subagent EngineRole: Coder Subagent

Best Paired With: Claude 3.7 Sonnet or GPT-5 as Lead Planner

Strong coding benchmarks allow running 4-8 parallel Coder workers across separate worktrees simultaneously.

Git Worktree Isolation Advantage:

Executes isolated bug fixes and author tests inside dedicated worktrees without cross-worker contention.

What Developers & Practitioners Are Saying

Real-world field observations, production strengths, and developer consensus

Field Consensus

"Next-gen compact model from OpenAI, delivering near-GPT-4o coding capabilities at mini-tier economics."

Verified Real-World Strengths

54.0% on SWE-bench Verified
Fast throughput
Exceptional tool orchestration for high-volume pipelines

Production Caveats & Gaps

Best paired with a primary planner for complex architectural tasks

Primary Sources & Methodology Citations

OpenAI GPT-5-mini Specifications — OpenAI (August 2025)

GPT 5 Mini model overview

GPT 5 Mini is listed in the public Swarm catalog through OpenAI. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.

Provider: OpenAI
Family: Gpt 5
Context window: 400K tokens
Thinking support: Not supported
Availability status: Available
Catalog ID: openai/gpt-5-mini

File input
Image input
Text input
Text output
Reasoning
Tools
Structured output
JSON
Batch

Run GPT 5 Mini in Swarm

Verified CLI workflow & multi-agent architecture

Quickstart Guide →

bash — swarm run

$swarm ctl model set --provider openai --model openai/gpt-5-mini

# Or switch interactively inside the terminal TUI:

>/models

Subagent Role Matrix

Assign GPT 5 Mini to specialized system roles in ~/.swarm/config.json:

Primary / Planner: Orchestrates full task DAGs and multi-stage pipelines.
Coder Subagent: Generates code and unit tests inside private Git worktrees.
Finder Subagent: Performs read-only code search, symbol discovery, and mapping.

System Agent Roles →Why Swarm Architecture →