OpenRouter model
Meta: Llama 3.3 70B Instruct
The Meta Llama 3.3 multilingual large language model (LLM) is a pretrained and instruction tuned generative model in 70B (text in/text out). The Llama 3.3 instruction tuned text only model...
- Provider
- OpenRouter
- Availability
- Available
- Model ID
- meta-llama/llama-3.3-70b-instruct
- Snapshot
- 2026-10-07
Meta: Llama 3.3 70B Instruct Benchmarks & Agentic Accuracy
Verified primary research & developer preference metrics
What This Means in Real Software Development
- Matches older 405B capabilities in a compact 70B footprint, offering strong efficiency.
- Scores 39.8% on SWE-bench Verified and 38.0% on LiveCodeBench.
- Full open weights permit local deployment, private fine-tuning, and on-premise execution.
- Widely available across cloud providers (Together, Fireworks, Groq, OpenRouter).
Swarm Multi-Agent Topology
Best Paired With: Claude 3.7 Sonnet as Lead Planner
Strong general code synthesis and reliable open-weights licensing for enterprise workflows.
Executes isolated feature branches inside Swarm worktrees.
What Developers & Practitioners Are Saying
Real-world field observations, production strengths, and developer consensus
"Meta's flagship open-weights workhorse. Extremely reliable instruction follower with broad industry fine-tuning support."
Verified Real-World Strengths
- High stability across diverse coding tasks and frameworks
- Solid 39.8% on SWE-bench Verified
- Broad provider availability across Fireworks, Together, Groq, and AWS Bedrock
Production Caveats & Gaps
- Coding performance slightly trails specialized coding models like Qwen 2.5 Coder
- Prone to standard conversational preamble unless strictly suppressed
Primary Sources & Methodology Citations
- Llama 3.3: Advancing Open Intelligence — Meta AI (December 2024)
Meta: Llama 3.3 70B Instruct model overview
Meta: Llama 3.3 70B Instruct is listed in the public Swarm catalog through OpenRouter. This page summarizes the verified catalog fields used by Swarm without claiming real-time access for every provider account.
- Provider
- OpenRouter
- Family
- OpenRouter
- Context window
- 131.1K tokens
- Thinking support
- Not supported
- Availability status
- Available
- Catalog ID
- openrouter/meta-llama/llama-3.3-70b-instruct
Run Meta: Llama 3.3 70B Instruct in Swarm
Verified CLI workflow & multi-agent architecture
swarm ctl model set --provider openrouter --model openrouter/meta-llama/llama-3.3-70b-instruct# Or switch interactively inside the terminal TUI:
/modelsSubagent Role Matrix
Assign Meta: Llama 3.3 70B Instruct 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.
Public catalog snapshot. Catalog availability is not a real-time provider status or an access guarantee.