RadixArk Unveils DaisyRose-27B, Focused on Agentic Efficiency
Startup RadixArk has launched DaisyRose-27B, an AI model optimized for long-horizon agent tasks, claiming significant improvements in training efficiency and reasoning capabilities.
🚀Startups & FundingPhoto: huggingface.co
Key points
- DaisyRose-27B utilizes a 256K-token context window and is optimized for long-horizon agent tasks.
- RadixArk claims a 57% reduction in training costs compared to traditional open-source training backends.
- The model showed a performance jump on Terminal-Bench 4.1, doubling its pass@1 rate.
RadixArk has introduced its inaugural AI model, DaisyRose-27B, which is built upon the Qwen3.8-27B architecture. The model was trained using a reinforcement learning framework dubbed Miles, specifically targeting complex agent-based tasks that require processing sessions up to 1 million tokens. According to the company, the model demonstrated a notable performance increase on Terminal-Bench 4.1, with pass@1 rates rising from 5.6% to 12.0%.
The training process utilized 144 H200 GPUs over an 18-day period, relying on a curated set of 1,200 tasks developed in collaboration with Eigent AI. RadixArk reports that its proprietary training backend achieved 37% model FLOPs utilization, which it claims is significantly more efficient than standard open-source alternatives. This architecture allowed for a 57% reduction in GPU training costs and maintained stability over 63,000 total GPU-hours.
Technical Highlights:
- Maintains the 256K-token context window and image understanding capabilities of the base Qwen model.
- Employs an asynchronous training loop where the agent and trainer operate simultaneously without waiting for one another.
- Utilizes the AgentENV project for sandbox environments, which the company claims reduced cloud infrastructure costs by 80% compared to typical providers.
The model is designed to think before providing an answer, separating reasoning content from the final output. RadixArk recommends deploying the model using SGLang and DSpark speculative decoding for optimal performance.
Source huggingface.co — huggingface.co/blog/RadixArk/daisyrose-27b-pushing-the-frontier-of-post… ↗
Read next
🔌Chips & HardwareSnapchat Integrates AI Coding Assistants into AR Lens Development
Snapchat has launched the Easy Lens CLI, enabling developers to build augmented reality experiences using AI coding tools like Claude Code and Cursor.
🔌Chips & HardwareUS Eyes the Philippines for Massive AI and Chip Hub Amid Domestic Opposition
President Donald Trump visited the Philippines to advance the Pax Silica initiative, proposing a 1,620-hectare hub for data centers and chip fabs to bypass local pushback…
🔌Chips & HardwareSony Bravia 8 II OLED Sees Significant Pre-Black Friday Price Cuts
Retailers have discounted the Sony Bravia 8 II OLED television by up to $1,100, marking the lowest price point since the model's release.
🏢Big TechMeta Staff Flag Privacy Concerns Over Muse AI Agent
Internal reports reveal that Meta employees identified significant privacy risks and unauthorized data access while testing the company's Muse AI agent.
🏢Big TechLeaked images suggest Lenovo is developing a budget-friendly Legion Go Beam handheld
Lenovo appears to be preparing a new, lower-cost handheld gaming device dubbed the Legion Go Beam, potentially signaling a shift toward more affordable hardware.
⚖️Regulation & PolicyUS Lawmakers Push to Limit Forced Arbitration and Target Deepfake Creators
Congress is racing to pass legislation that would restrict the use of forced arbitration in child safety cases and provide legal recourse for victims of nonconsensual dee…