DaisyRose-27B: 1,200-task RL took Qwen3.8 from 5.6% to 12% on Terminal-Bench 4.1
Comments on “DaisyRose-27B: 1,200-task RL took Qwen3.8 from 5.6% to 12% on Terminal-Bench 4.1”
3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked
RadixArk released DaisyRose-27B, a version of Alibaba's Qwen3.8-27B fine-tuned with reinforcement learning on 1,200 curated long-horizon agent tasks over 18 days on 144 H200 GPUs.
Why it ranked
No. 102 of 144 with a score of 0.59. The panel rated it highest on new information (0.88) and technical depth (0.75), lowest on significance (0.54). The judges disagreed most on significance (spread 0.42), so the story is marked contested. Placed No. 102, No. 99 by score alone: the 43rd AI & machine learning story of the day (−0.080). The placement terms move stories; they are not part of the score.
- AI judgeJudge 2claude-haiku-5-5
- Model release with RL training details and infrastructure cost claims; useful engineering detail and open checkpoints, but benchmarks are self-reported, gains modest, and some promotional framing.}
- AI judgeJudge 3gpt-5.6-luna
- A substantive model release combines reproducible checkpoints, benchmark methodology, RL task design, training efficiency, and sandbox infrastructure metrics. It is highly relevant to agent and model builders, though results are primarily self-reported and the post has some launch-oriented framing.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.59total under rubric v2.3Contestedjudges differ by 0.42 on significance
- Significance0.54
- Reader interest0.74
- Audience breadth0.56
- Practical relevance0.68
- Substantiated0.69
- Technical depth0.75
- New information0.88
- Counts against: Promotional0.22
- Counts against: Overclaim0.30
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.60 | 0.30 | 0.72 | 0.54 | 0.42 , above the contested threshold | +0.24 | +0.130 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.68 | 0.60 | 0.93 | 0.74 | 0.33 | +0.26 | +0.192 |
| Audience breadthHow many people who build technology does it concern? | 0.61 | 0.40 | 0.68 | 0.56 | 0.28 | +0.13 | +0.073 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.71 | 0.55 | 0.78 | 0.68 | 0.23 | +0.22 | +0.150 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.82 | 0.45 | 0.80 | 0.69 | 0.37 | +0.05 | +0.034 |
| Technical depthHow technically substantive is it? | 0.82 | 0.60 | 0.82 | 0.75 | 0.22 | +0.05 | +0.037 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.85 | 0.80 | 0.98 | 0.88 | 0.18 | +0.05 | +0.044 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.12 | 0.35 | 0.18 | 0.22 | 0.23 | −0.19 | −0.041 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.29 | 0.40 | 0.20 | 0.30 | 0.20 | −0.11 | −0.033 |
| Sum of contributions | +0.586 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.586 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-11-19/102/"><img src="https://rubric.hrih.si/badge/2026-11-19/102.svg" alt="Ranked #102 on Rubric, 19 Nov 2026" width="252" height="28"></a>Also covered by
Elsewhere
Elsewhere: HN 172 pts, front page
Recorded for context only, checked 19 Nov 2026, 05:47 UTC. It played no part in the ranking.