swarmagent.dev
GitHub
DocsModelsResourcesPricingNewsContactGitHub

Fireworks AI model

OpenAI gpt-oss-120b

Welcome to the gpt-oss series, OpenAI's open-weight models designed for powerful reasoning, agentic tasks, and versatile developer use cases. gpt-oss-120b is used for production, general purpose, high reasoning use-cases that fits into a single H100 GPU.

Provider
Fireworks AI
Availability
Available
Model ID
gpt-oss-120b
Snapshot
2026-10-07
Evaluation Verdict

OpenAI gpt-oss-120b is an open-weights coding and reasoning model (55.4% SWE-bench Verified, 74.0% AIME, 1350 Coding Elo), providing enterprise-grade local execution without cloud dependency.

OpenAI gpt-oss-120b Benchmarks & Agentic Accuracy

Verified primary research & developer preference metrics

Primary Sources Verified
SWE-bench Verified
55.4%
StandardOpenAI Open Weights Report ↗
LiveCodeBench
51.2%
StandardLiveCodeBench ↗
LMSYS Coding Elo
1350Elo
Coding ArenaLMSYS Arena ↗
AIME 2024 / 2025
74%
StandardOpenAI Open Weights Report ↗
Pricing ($ / 1M tokens)
$0.4/ $1.2
$0.08 cached readStandard API Pricing

What This Means in Real Software Development

  • Open-weights frontier architecture capable of running on private enterprise GPU clusters.
  • Scores 55.4% on SWE-bench Verified, matching or exceeding many proprietary cloud models.
  • Fast inference on modern accelerator hardware (NVIDIA H100/B200).
  • 100% private and air-gapped execution for sensitive codebases.

Swarm Multi-Agent Topology

Tier 2: Open-Weights Enterprise EngineRole: Coder Subagent

Best Paired With: Claude 3.7 Sonnet as Lead Planner

High code synthesis fidelity for teams requiring on-premise execution with zero external data transfer.

Git Worktree Isolation Advantage:

Executes inside local Swarm worktrees via vLLM or local OpenAI-compatible endpoints.

What Developers & Practitioners Are Saying

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

Field Consensus

"OpenAI's open-weights enterprise foundation model, offering transparent weights for private cloud and sovereign deployments."

Verified Real-World Strengths

  • 55.4% on SWE-bench Verified
  • Complete self-hosting capability with full data privacy
  • Broad multi-language programming competence

Production Caveats & Gaps

  • Requires multi-GPU infrastructure (4x A100/H100) to serve at scale
Benchmark Reality Check

What the benchmark tests: Tests enterprise software development on open-weights foundation models.

Production caveat: Deploy on private GPU clusters for zero-exfiltration code workflows.

Primary Sources & Methodology Citations

OpenAI gpt-oss-120b model overview

OpenAI gpt-oss-120b is listed in the public Swarm catalog through Fireworks AI. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.

Provider
Fireworks AI
Family
Gpt Oss
Context window
131.1K tokens
Thinking support
low · medium · high
Availability status
Available
Catalog ID
fireworks/gpt-oss-120b
  • Text input
  • Text output
  • Reasoning
  • Tools

Run OpenAI gpt-oss-120b in Swarm

Verified CLI workflow & multi-agent architecture

Quickstart Guide →
bash — swarm run
$swarm ctl model set --provider fireworks --model fireworks/gpt-oss-120b

# Or switch interactively inside the terminal TUI:

>/models

Subagent Role Matrix

Assign OpenAI gpt-oss-120b 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 →
← All Fireworks AI models

Public catalog snapshot. Catalog availability is not a real-time provider status or an access guarantee.