Hybrid-Phi-Mamba-1.5B
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
- Carnegie Mellon University (CMU),Mohamed bin Zayed University of Artificial Intelligence (MBZUAI),Cartesia
- Organisation type
- Academia,Academia,Industry
- Country
- United States of America, United Arab Emirates
- Published
- 19 August 2024
- Authors
- Aviv Bick, Kevin Y. Li, Eric P. Xing, J. Zico Kolter, Albert Gu
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling/generation
- Base model
- Phi-1.5
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
- 1.5B
- Training data
- 5,000,000,000 tokens
"We distill the Phi-1.5-1.3B model into Phi-Mamba1.5B as well as Hybrid-Phi-Mamba-1.5B" (https://arxiv.org/pdf/2408.10189, page 8).
"The Hybrid-PhiMamba-1.5B is distilled on a budget of 5 billion tokens from the same dataset."
Training compute
The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.
- Training compute
- 1.2 × 10²¹ FLOP
- How it was established
- Operation counting
- Fine-tuning compute
- 4.5 × 10¹⁹ FLOP
1.17e+21 FLOP [base model compute] + 4.5e+19 FLOP [finetune compute] = 1.215e+21 FLOP
6 FLOP / parameter / token * 1.5 * 10^9 parameters * 5 * 10^9 tokens = 4.5e+19 FLOP
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
- Unreleased
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
- Transformers to SSMs: Distilling Quadratic Knowledge to Subquadratic Models
- Last updated
- 28 November 2025
What the numbers mean
Where it came from
Hybrid-Phi-Mamba-1.5B was published by Carnegie Mellon University (CMU),Mohamed bin Zayed University of Artificial Intelligence (MBZUAI),Cartesia, in United States of America, in August 2024. academia,Academia,Industry is the category the publisher falls under.
It works in Language, and is recorded as doing language modeling/generation.
Its starting point was Phi-1.5 — most models at this scale are adapted from an existing base rather than built from nothing.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
Training and provenance
Training it took roughly 1.2 × 10²¹ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.
The training set ran to roughly 5,000,000,000 tokens.
Answers
Hybrid-Phi-Mamba-1.5B — common questions
How much compute was used to train Hybrid-Phi-Mamba-1.5B?
Around 1.2 × 10²¹ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.
What GPU do I need to run Hybrid-Phi-Mamba-1.5B?
None. Hybrid-Phi-Mamba-1.5B 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 Hybrid-Phi-Mamba-1.5B open source?
No. Hybrid-Phi-Mamba-1.5B has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does Hybrid-Phi-Mamba-1.5B have?
Hybrid-Phi-Mamba-1.5B has 1.5B parameters. "We distill the Phi-1.5-1.3B model into Phi-Mamba1.5B as well as Hybrid-Phi-Mamba-1.5B" (https://arxiv.org/pdf/2408.10189, page 8). 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 Hybrid-Phi-Mamba-1.5B?
Hybrid-Phi-Mamba-1.5B was published by Carnegie Mellon University (CMU),Mohamed bin Zayed University of Artificial Intelligence (MBZUAI),Cartesia, based in United States of America, categorised as academia,Academia,Industry.
When was Hybrid-Phi-Mamba-1.5B released?
Hybrid-Phi-Mamba-1.5B was published in August 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is Hybrid-Phi-Mamba-1.5B used for?
Hybrid-Phi-Mamba-1.5B works in Language, and is recorded as handling language modeling/generation. These are the areas it was designed around; they describe intent rather than a hard boundary.
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.