LoongRL 14B
No estimate
No hardware requirements for this model
The weights for this model have not been published, so it cannot be downloaded or run on your own hardware at any size. It is reachable only through its provider, and no graphics card changes that.
On record
Full specification
Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.
Origin
Who built this model, where, and when it was published.
- Organisation
- Microsoft Research Asia,Shanghai Jiao Tong University,Carnegie Mellon University (CMU)
- Organisation type
- Industry,Academia,Academia
- Country
- China, United States of America
- Published
- 27 October 2025
- Authors
- Siyuan Wang, Gaokai Zhang, Li Lyna Zhang, Ning Shang, Fan Yang, Dongyao Chen, Mao Yang
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling/generation, Question answering, Quantitative reasoning
- Base model
- Qwen2.5-14B
Size
How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.
- Parameters
- 14B
- Training data
- tokens
14B
"Batch sizes are set to 512 for 7B model and 256 for 14B model" "a maximum output length of 4,096 tokens and longcontext inputs of ∼ 16K" "We train for 168 steps in Stage I and 150 steps in Stage II" "After Stage I, we generate eight rollouts per example using the best checkpoint"
The training run
What it physically took to train: which chips, how many, for how long, and what that drew from the wall.
- Training hardware
- AMD Instinct MI300X
- Chips used
- 8
- Power draw
- 11.7 kW
Availability
Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.
- Weights
- Closed — provider access only
- Model access
- Unreleased
- Training code
- Open (non-commercial)
"We have made extensive efforts to ensure the reproducibility of our work. Details of the GRPO algorithm and hyperparameters are provided in Section 3.2.1 and Section 4.1. We provide our training prompt template in Appendix A.2. The datasets used in our experiments are described in Table 1. To further facilitate reproducibility, the supplementary materials include (i) our RL training code, (ii) the code for synthesizing KeyChain data, and (iii) several representative samples of the synthesized K…
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Confident
Sources
Where this record came from and when it was last checked.
- Reference
- LoongRL: Reinforcement Learning for Advanced Reasoning over Long Contexts
- Last updated
- 28 November 2025
What the numbers mean
Where it came from
LoongRL 14B was published by Microsoft Research Asia,Shanghai Jiao Tong University,Carnegie Mellon University (CMU), in China, in October 2025. The organisation is categorised as industry,Academia,Academia.
It works in Language, and is recorded as doing language modeling/generation, Question answering, Quantitative reasoning.
It builds on Qwen2.5-14B, which is why it shares that model's general shape and size.
Its weights were never published, so it can only be reached through its provider. No graphics card changes that.
Answers
LoongRL 14B — common questions
What is LoongRL 14B used for?
LoongRL 14B works in Language, and is recorded as handling language modeling/generation, Question answering, Quantitative reasoning. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
What GPU do I need to run LoongRL 14B?
None. LoongRL 14B is a closed model — its weights were never published, so it cannot be downloaded or run on your own hardware at any price. It is reachable only through its provider.
Is LoongRL 14B open source?
No. LoongRL 14B has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does LoongRL 14B have?
LoongRL 14B has 14B parameters. 14B. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.
Who created LoongRL 14B?
LoongRL 14B was published by Microsoft Research Asia,Shanghai Jiao Tong University,Carnegie Mellon University (CMU), based in China, categorised as industry,Academia,Academia.
When was LoongRL 14B released?
LoongRL 14B was published in October 2025.
The other direction
Looking at it from the other side?
This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.