PulseAugur
EN
LIVE 10:02:07

Hark's Handoff² agent achieves 2x speedup using RL

Hark has released Handoff², an updated version of its computer-use agent for Hark Pro, which is reportedly twice as fast as its predecessor. The speed improvements are attributed to reinforcement learning techniques that penalize wasted turns and latency, leading the agent to batch actions and reuse composite actions. While Handoff² achieved 82% task completion on Hark's internal benchmark, it still trails behind Claude Opus 5.5 in certain benchmarks. AI

IMPACT This update demonstrates advancements in agent speed and efficiency through reinforcement learning, potentially influencing the development of other browser-based AI agents.

RANK_REASON Product update for an AI-adjacent tool, not a core AI model release.

Read on MarkTechPost →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Hark's Handoff² agent achieves 2x speedup using RL

How we ranked this

Signal score
36 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Product update for an AI-adjacent tool, not a core AI model release.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
product, infra
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. MarkTechPost TIER_1 English(EN) · Michal Sutter ·

    Hark Releases Handoff², a 2x Faster Computer-Use Agent for Hark Pro

    <p>Hark has released Handoff², the computer-use agent inside Hark Pro. It averages 2x faster than v1, completes 82% of 478 internal web tasks, and learned to batch actions through latency-penalized RL.</p> <p>The post <a href="https://www.marktechpost.com/2026/11/16/hark-handoff-…