Skip to content

RadixArk opens DaisyRose-27B, an agent-tuned Qwen model

Short answer

RadixArk released DaisyRose-27B, its first model, as open weights under Apache 2.0. It is Qwen3.8-27B further trained with reinforcement learning on long agent tasks. Terminal Bench 4.1 pass@1 moved from 5.6% to 12.0%. The scores are the vendor's own, so test it on your tasks.

What this means for operators

If you run sales, support or operations at a 10-200 person company and have been testing agents that execute multi-step work, this is a model you can self-host and trial without a licence negotiation. It exposes an OpenAI-style endpoint through SGLang, so a pilot can reuse client code you already have. The absolute numbers are modest: 12.0% on the hardest terminal benchmark means most long tasks still fail. Run it on a handful of your real workflows, such as report generation or data clean-up, count completed jobs, and compare with the base Qwen3.8-27B before committing any GPU budget.

RadixArk has published DaisyRose-27B on Hugging Face, the company's first model. It builds on Qwen3.8-27B and was trained with reinforcement learning to handle long-running agent work, with sessions of up to 1M tokens.

What is released

Checkpoints are available in BF16, FP8 and NVFP4 formats under the Apache 2.0 licence. The model keeps the base model's architecture, tokenizer, chat template, image understanding and 256K-token context window. Longer training sessions were stitched together from several trajectories through summarization by the agent harness.

RadixArk recommends serving it with SGLang plus DSpark speculative decoding, and points to an SGLang cookbook with launch settings for different GPUs and checkpoint formats. Once the server is running, it can be called with the standard OpenAI Python client. The model reasons before answering by default, and returns its reasoning and final answer in separate fields.

Reported results

According to RadixArk, DaisyRose-27B beats Qwen3.8-27B on every benchmark it lists:

  • Terminal Bench 4.1: 5.6 to 12.0
  • Terminal Bench 2.1: 73.0 to 75.5
  • DeepSWE v1.1: 42.2 to 46.2
  • SWE Bench Pro v2: 91.3 to 94.7
  • IFBench: 79.5 to 82.4
  • GPQA Diamond: 89.2 to 90.7
  • HLE (text-only): 33.9 to 34.3

The company reports the higher of the BF16 and FP8 runs per benchmark, and says the two differ by under one point where both were measured. These are the vendor's own figures; no independent evaluation is cited.

How it was trained

The training set was about 1,200 tasks, built with Eigent AI from thousands of candidates. Each task was redesigned around a mistake the base model makes, requires several dependent steps, and was reviewed so it cannot be passed without solving the actual problem. Each attempt gets up to 1.5 hours in a sandbox.

Training ran for 18 days on 144 H200 GPUs, 128 for running agents and 16 for the trainer, in a fully asynchronous setup. RadixArk says its internal backend reached 37% model FLOPs utilization against 16% for the open-source trainer, which it translates into 57% less GPU time. The run totalled 63,000 GPU-hours.

The sandbox service ran about 944,000 sandboxes between mid-September and mid-November, up to 1,566 at once. RadixArk reports a creation success rate above 99%, and cloud costs about 80% below popular sandbox providers. The training backend and sandbox service figures describe RadixArk's own infrastructure; only the model weights are the released artifact in this announcement.

What to check

RadixArk asks users to report results on their own tasks, including failures, in the model's Community tab. That is the right test: the gains are measured on coding, terminal and reasoning benchmarks, and the source says nothing about sales, support or back-office workloads.

Source: Hugging Face · In the Atlas: Hugging Face →

The tools in this story

  • OpenAI - price, who uses it, alternatives
Next step

Visibility Analyzer

An AI news item will not tell you how assistants see your own site. The Visibility Analyzer checks that on your live site: which answers cite you, which pages an engine cannot retrieve, and what to fix first. Free to run.

Run a free visibility audit

Free to run. No card.

Fee
Free
Length
One run, minutes

Free tier: two analyses a day, no card required.