Skip to main content
Kaino.dev
Discover
Evals
News
Academics
Insights
Kaino.dev

Discover, evaluate, and compare AI tools, models, and agents.

Explore

  • Discover
  • Evaluations
  • News
  • Academics
  • Insights

Community

  • Twitter
  • YouTube
  • Instagram
Privacy PolicyTerms of Service

© 2026 Kaino.dev. All rights reserved.

Version 1.1.0
OpenHands LM · Discover · Kaino
Discover/MODELS/OpenHands LM
OpenHands LM logo

MODELS

OpenHands LM

by All Hands AI / OpenHands

modelagentic-codingOpenHands
Visit WebsiteDocumentationGitHub

Overview

OpenHands LM 32B v0.1 is an open coding-agent model from OpenHands / All Hands AI for autonomous software-development tasks.

Details

OpenHands LM 32B v0.1 is described in the OpenHands launch post as a strong, open coding-agent model built on Qwen Coder 2.5 Instruct 32B. The launch post reports a 128K context window and a 37.2% SWE-Bench Verified resolve rate. The official Hugging Face model card lists it as a 33B-parameter BF16 safetensors text-generation model downloadable for local use, and OpenRouter lists it as all-hands/openhands-lm-32b-v0.1 with a 131K context limit.

When to Use

Use when evaluating open coding models optimized for autonomous software-development agents. Use with OpenHands workflows when you want a model family documented by the OpenHands project. Use for local experimentation when the Hugging Face model card and MIT license meet your deployment requirements.

Getting Started

  1. Read the OpenHands LM launch post for the model’s positioning
  2. benchmark claim
  3. base model
  4. and context-window details.
  5. Review the Hugging Face model card at OpenHands/openhands-lm-32b-v0.1 to download the weights and confirm license and file format.
  6. Follow the OpenHands LLM provider documentation to configure CLI or SDK usage with an OpenHands LLM API key and the openhands/ model prefix.
  7. If using OpenHands for agent execution
  8. review the runtime architecture docs for the Docker-based sandbox model.

Key Features

  • •32B open coding-agent model introduced by OpenHands / All Hands AI.
  • •Built on Qwen Coder 2.5 Instruct 32B according to the launch post.
  • •Launch post reports a 128K context window and 37.2% SWE-Bench Verified resolve rate.
  • •Hugging Face model card lists 33B parameters
  • •BF16 safetensors
  • •and text-generation support.
  • •OpenRouter catalog page lists the model as all-hands/openhands-lm-32b-v0.1 and describes it as optimized for autonomous software-development agents.

Capabilities

  • •text-generation
  • •coding
  • •software-engineering-agent-support
  • •long-context

Last updated Jul 23, 2026