Primary sources last checked on 19 Nov 2026. Verification freshness is tracked separately from factual change, and unknown values remain explicit rather than inferred.
NotableOfficial publisher original checkpoint is publicly accessible at an immutable 40-character revision, with concrete visible weights and a source-specified license. Independent OMW runtime or benchmark performance has not been established; Notable only.Curation method →
Open Model Weights records DaisyRose-27B as a 27.8B including vision encoder and MTP head (publisher)-parameter multimodal image/text to text, agent workflows open-weight model in the DaisyRose family, with RadixArk listed as the developer. The checked source repository provides 19 recognized weight artifacts, format evidence for Safetensors and a verified context window of 262,144 tokens, publisher-declared; not independently tested. Values that are not evidenced in the checked sources remain explicitly undisclosed or unknown.
The repository declares APACHE-2.0. Commercial use allowed under apache-2.0 license conditions. Open Model Weights reports the checked license evidence and classification; this is not legal advice.
The checked sources declare Qwen/Qwen3.8-27B as lineage evidence. Training recipe status: publisher-described-not-reproduced; data disclosure status: Publisher describes task selection and reward process; source dataset is not fully verified as independently downloadable.
DaisyRose-27B is RadixArk's openly released reinforcement-learning post-trained derivative of Qwen3.8-27B, not a newly pretrained foundation model. It retains 262,144-token context and multimodal input pathways; the official BF16 checkpoint exposes 18 safetensors shards and an index under Apache-2.0. The model profile links an exact original-repository revision and visible files; hardware and benchmark performance remain untested by Open Model Weights.
Official publisher sourceRadixArk/DaisyRose-27BSource ↗
Original weight evidence19 explicitly recorded source-visible Safetensors path(s)Source ↗
License provenanceapache-2.0 — original LICENSE fileSource ↗
Parameter and context claim27.8B including vision encoder and MTP head (publisher); 262,144 tokens, publisher-declared; not independently testedSource ↗
Benchmark statusPublisher or external leaderboard claims only; no independent OMW inference or benchmarkSource ↗
WHY IT MATTERS
The official source card describes long-horizon RL training designed to increase terminal and coding-agent task completion rather than changing the pretrained architecture.
Eighteen numbered weight shards plus model.safetensors.index.json are visible in the source repository at a pinned commit, with a standard Apache 2.0 LICENSE file.
Terminal-Bench 4.1 figures are developer-run harness comparisons, not independently verified OMW testing; the repository also points to separate FP8 and NVFP4 derivatives.
MODEL-SELECTION VIEW
Strong fit when…
Comparing long-horizon coding agents with pinned Qwen-family baselines.
Studying tool-use and terminal behavior after reinforcement-learning post-training.
Serving source-verified 27B multimodal agents with compatible runtimes.
Important considerations
A training trajectory extending past 256K through summarization does not mean a one-million-token inference context.
Published benchmark gains depend on a particular harness and task set.
FP8 and NVFP4 are separate repositories and should not be silently treated as part of the BF16 file inventory.
Model purpose and architecture
DaisyRose-27B is RadixArk's openly released reinforcement-learning post-trained derivative of Qwen3.8-27B, not a newly pretrained foundation model. It retains 262,144-token context and multimodal input pathways; the official BF16 checkpoint exposes 18 safetensors shards and an index under Apache-2.0.
The official source card describes long-horizon RL training designed to increase terminal and coding-agent task completion rather than changing the pretrained architecture.
Pinned public weight artifacts
Eighteen numbered weight shards plus model.safetensors.index.json are visible in the source repository at a pinned commit, with a standard Apache 2.0 LICENSE file.
The checked immutable source revision is 293a7a2ab54aaea78098ee917c546390667521dc. All recorded primary weight paths are from the official pinned repository tree. A file name is not a measured runtime result.
License and provenance
Terminal-Bench 4.1 figures are developer-run harness comparisons, not independently verified OMW testing; the repository also points to separate FP8 and NVFP4 derivatives.
License evidence: a standard original LICENSE file under apache-2.0. The checkpoint license does not automatically cover third-party datasets, application code or generated outputs.
Performance evidence and limitations
These are publisher or independently hosted third-party benchmark claims, not tests Open Model Weights has executed. A fair comparison requires an exact checkpoint, matched prompts, harness, hardware and repeatable samples.
A training trajectory extending past 256K through summarization does not mean a one-million-token inference context.
Deployment and practical evaluation
Appropriate evaluations include: Comparing long-horizon coding agents with pinned Qwen-family baselines. Studying tool-use and terminal behavior after reinforcement-learning post-training. Serving source-verified 27B multimodal agents with compatible runtimes. Source code and quantizations from separate repositories must be checked independently.
Published benchmark gains depend on a particular harness and task set. FP8 and NVFP4 are separate repositories and should not be silently treated as part of the BF16 file inventory.
IMPLEMENTATION & EVALUATION
How to evaluate DaisyRose-27B before deployment
Primary-source pinned checkpoint evaluation steps; no independent OMW model inference performed.
STEP 01
Pin official checkpoint and artifact paths
Publisher repositories can change after publication.
REPRODUCIBLE ACCEPTANCE TEST
Record the source 40-character revision and verified visible filenames, along with processor/config revision.
WHAT CAN GO WRONG
Mutable tags, missing shards and silent upstream changes.
STEP 02
Review license and code dependencies
Open model weights do not imply unrestricted use of code or all data.
REPRODUCIBLE ACCEPTANCE TEST
Inspect the exact publisher license source, custom modeling code, dependencies and intended use policy.
WHAT CAN GO WRONG
Misclassified commercial license or remote-code execution.
STEP 03
Measure hardware, context and workload
Advertised parameters, context and token rates are not observed system requirements.
REPRODUCIBLE ACCEPTANCE TEST
Publish repeatable measured memory, batch, hardware, latency and context settings.
WHAT CAN GO WRONG
Out-of-memory, unsupported attention operators and overclaimed throughput.
STEP 04
Separate external claims from OMW results
Publisher comparisons have different harnesses.
REPRODUCIBLE ACCEPTANCE TEST
Use held-out tasks and matched baselines before independent ranking.
WHAT CAN GO WRONG
False benchmarking attribution or ungrounded Leading status.
Are the original DaisyRose-27B model weights publicly available?
Yes. The official source repository RadixArk/DaisyRose-27B exposes actual Safetensors checkpoint artifacts at the pinned revision 293a7a2ab54aaea78098ee917c546390667521dc.
Is DaisyRose-27B independently benchmarked by Open Model Weights?
No. Published scores and throughput descriptions are attributed to their source; OMW has not run model inference or a controlled benchmark.
What license applies to this exact checkpoint?
apache-2.0. Evidence is the repository's original license file; follow exact terms and separately inspect dependencies.
Is the model suitable for local deployment?
Original weights are downloadable, but actual fit, operators and performance need hardware-specific testing. Published benchmark gains depend on a particular harness and task set.
Does Notable mean the model leads an independent benchmark?
No. Notable records a distinct useful, source-verifiable original checkpoint. Leading requires separate matched independent evidence.
03 · PRACTICAL TECHNICAL PROFILE
Deployment, licensing and reproducible evidence.
Inspect weight-memory estimates, model lineage and runtime evidence separately from promotional or untested performance claims.
DEPLOYMENT PROFILE
Weight memory, without pretending it is a deployment guarantee.
These bars visualize weight-only decimal-GB estimates. They exclude KV cache, activations, runtime overhead, optimizer state and sharding.
WEIGHT-ONLY MEMORY
BF16 / FP16~55.6 GB
FP8 / INT8~27.8 GB
INT4~13.9 GB
FIELD EVIDENCE MAP
WeightsVerified
LicenseVerified
Commercial useAllowed
ContextNot declared
FormatsVerified
LineageDeclared
Training recipePublisher described not reproduced
Training dataPublisher describes task selection and reward process; source dataset is not fully verified as independently downloadable
RuntimeSource-derived · not runtime-tested
LICENSE & USE
APACHE-2.0
Commercial use allowed under apache-2.0 license conditions
Declared licenseAPACHE-2.0
ClassificationAllowed
VerificationLicense file checked
This section summarizes the checked repository evidence. It is not legal advice. Inspect license evidence ↗
LINEAGE & PROVENANCE
What the checked sources disclose — and what they do not.
Open Model Weights does not infer a hidden base model or training corpus when the checked sources do not declare one.
LINEAGE
BASE / PARENTQwen/Qwen3.8-27BDeclared
fine-tuned-from-or-based-on
CURRENT RECORDDaisyRose-27BDaisyRose
TRAINING DISCLOSURE
Training recipePublisher described not reproduced
Training dataPublisher describes task selection and reward process; source dataset is not fully verified as independently downloadable
FamilyDaisyRose
RUNTIME SIGNALS
Compatibility claims stay attached to their evidence level.
Declared and model-card-mentioned runtimes are discovery signals. They are not treated as Open Model Weights runtime tests unless explicitly measured.
MentionedTransformers
MentionedSGLang
MentionedvLLM
OMW runtime testNot performed
EVIDENCE TRAIL
Sources and observed history.
This is where the article remains auditable: source links on one side, the registry's observed change history on the other.
Three evidence-linked visual summaries: parameter scale, illustrative precision memory and verifiable provenance. No performance ranking is implied.
01 / PARAMETER SCALE
Total versus active parameters
Total parameters27.8B
Active parameter count not established.
Repository-derived or publisher-described parameter count
Published total parameters are not a model-quality or performance score.
02 / MEMORY ESTIMATE
Weight-only precision footprint
BF16 / FP1655.6 GB
FP8 / INT827.8 GB
INT413.9 GB
Hypothetical weight-only decimal GB, not measured peak VRAM or guaranteed support for each format. KV cache, activations, runtime and sharding are excluded.
Repository SHA 293a7a2ab54a…; a source check is not an inference benchmark or legal clearance. See the per-model evidence JSON for the exact source record.
A verification check does not pretend the model changed. Open Model Weights updates the verification date after a real primary-source check, while factual-update dates move only when retained source evidence changes.
Declared, mentioned and externally evidenced runtimes
Declared librarytransformers
Declared/taggedNone
Mentioned in model cardTransformers, SGLang, vLLM
Artifact signalsNone
External primary runtime evidenceNone
Tested by Open Model WeightsNo — source-derived support only
Open Model Weights separates model-repository evidence from external runtime-release evidence. An external support statement does not change the official model repository's weight formats and does not imply an Open Model Weights inference benchmark.
training recipesource-described-not-independently-reproduced
data disclosurePublisher describes task selection and reward process; source dataset is not fully verified as independently downloadable
runtime supportsource-derived-not-runtime-tested
MODELS FOR AI AGENTS · ROLE EVIDENCE
Where this model fits in an agent system
Specialized model. A supporting component. This checkpoint does not, by category alone, serve as the planning and tool-execution controller.
The curated notable tier is not an agent-readiness rating. Model-specific tool-call success, long-horizon task accuracy and compatible runtime configurations are not independently established here.
Source-linked license: APACHE-2.0 (allowed commercial-use classification; see original terms).
Context: 262144 configured tokens.
BF16 weights: 55.6 estimated decimal GB, not runtime VRAM.