LLM Reference

Holo1.5-3B

Released
2025-09-25
Last refreshed
2026-06-07
Status
Researched 44d ago
Open weightsCommercial use: conditionalMultimodalAgentsLong contextVisionAgentsVision

Holo1.5-3B is available now for agents, long context, and vision with open-weight and 128k context; evaluate it while provider pricing coverage matures.

Use it for

  • Teams evaluating agents, long context, and vision
  • Workloads that can use a 128k context window

Do not use it for

  • Cost-sensitive launches that need sourced token pricing
  • Strict JSON or tool-calling flows
  • Teams that need a tracked hosted API route today
Specifications
Family
Holo
Released
2025-09-25
Context
128k
Parameters
3
Specialization
computer-use
Openness
Open weights
License
Qwen LicenseCommercial use: conditional
Weights
Available
Code
Unknown
Training
Reinforcement Learning
Created by

Agentic AI for Action

Paris, France
Founded 2023
Website
Pricing

No tracked provider token pricing is available yet.

About

Holo1.5-3B is H Company's 3B-parameter computer-use VLM, part of the Holo1.5 family released September 25, 2025. Trained with SFT + GRPO reinforcement learning; natively supports up to 3840x2160 resolution. Achieves 72.81% average UI localization score across six benchmarks, competitive with prior 7B-class models.

Holo1.5-3B is H Company's 3B-parameter computer-use VLM, part of the Holo1.5 family released September 25, 2025. Trained with SFT + GRPO reinforcement learning; natively supports up to 3840x2160 resolution. Achieves 72.81% average UI localization score across six benchmarks, competitive with prior 7B-class models.

Holo1.5-3B has a 128k-token context window.

Top use-case fit: coding, agents, and build tasks

Agents

Included by capability and metadata signals in the decision map.

Long context

Included by capability and metadata signals in the decision map.

Vision

Included by capability and metadata signals in the decision map.

Provider price ladder

No tracked provider token pricing is available for this model yet.

Capabilities

VisionMultimodal

Benchmark peer barsfor Agents

No task-mapped benchmark peers are available for this model yet.

Benchmark scores(3)

Scores are benchmark-specific and are direction-aware: the same numeric gap can mean very different outcomes across suites. Use the leaderboard context and this model's provider route to decide whether the winning margin is meaningful for your workload.
BenchmarkScoreVersionEvaluationSource
WebClick81.5WebClick - UI localizationObserved 2025-09-25Source
ScreenSpot-Pro51.5ScreenSpot-Pro - UI localizationObserved 2025-09-25Source
OSWorld-G61.6OSWorld-G - UI groundingObserved 2025-09-25Source

Migration checks

No linked migration route is available for this model yet.

Frequently asked questions

What is the context window of Holo1.5-3B?

Holo1.5-3B has a context window of 128k tokens.

When was Holo1.5-3B released?

Holo1.5-3B was released on 2025-09-25.

What benchmarks has Holo1.5-3B been tested on?

Holo1.5-3B has been evaluated on 3 benchmarks, including WebClick, ScreenSpot-Pro, OSWorld-G.