Composer 2
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
- Cursor
- Organisation type
- Industry
- Country
- United States of America
- Published
- 19 March 2026
- Authors
- Aaron Chan, Ahmed Shalaby, Alexander Wettig, Aman Sanger, Andrew Zhai, Anurag Ajay, Ashvin Nair, Charlie Snell, Chen Lu, Chen Shen, Emily Jia, Federico Cassano, Hanpeng Liu, Haoyu Chen, Henry Wildermuth, Jacob Jackson, Janet Li, Jediah Katz, Jiajun Yao, Joey Hejna, Josh Warner, Julius Vering, Kevin Frans, Lee Danilek, Less Wright, Lujing Cen, Luke Melas-Kyriazi, Michael Truell, Michiel de Jong, Na…
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Coding
- Base model
- Kimi K2.5
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
- 1T
- Training data
- tokens
Based on Kimi K2.5; 1.04T total, 32B active parameters
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
- 2.3 × 10²⁵ FLOP
Said by Lee Robinson on X - https://x.com/leerob/status/2035035355364081694 - to be 4x the training compute of the base model, Kimi K2.5; 4 x 5.8e24 = 2.32e25. B300 speculated based on Colossus 2 confirmation (https://x.com/elonmusk/status/2056422097237283295).
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
- NVIDIA B300 (Blackwell Ultra)
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
- API access
- Training code
- Unreleased
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Why it is tracked
- Discretionary
- Record confidence
- Confident
Model available by default in Cursor, which has >1M DAU; reasonable to assume >1M MAU for this model.
Sources
Where this record came from and when it was last checked.
- Reference
- Composer 2 Technical Report
- Last updated
- 22 May 2026
What the numbers mean
What this model is
Composer 2 was published by Cursor, in United States of America, in March 2026. The organisation is categorised as industry.
It works in Language, and is recorded as doing coding.
It builds on Kimi K2.5, which is why it shares that model's general shape and size.
Because the weights are not available, none of the hardware figures elsewhere on this site apply to it.
How it was trained
The training run consumed about 2.3 × 10²⁵ FLOP, on NVIDIA B300 (Blackwell Ultra). That figure describes the cost of creating it and has no bearing on how quickly it generates text.
It is tracked in the underlying dataset for one reason in particular: discretionary.
Answers
Composer 2 — common questions
What GPU do I need to run Composer 2?
None. Composer 2 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 Composer 2 open source?
No. Composer 2 has not had its weights published, so it exists only as a service controlled by its owner.
How many parameters does Composer 2 have?
Composer 2 has 1T parameters. Based on Kimi K2.5; 1.04T total, 32B active parameters. 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 Composer 2?
Composer 2 was published by Cursor, based in United States of America, categorised as industry.
When was Composer 2 released?
Composer 2 was published in March 2026.
What is Composer 2 used for?
Composer 2 works in Language, and is recorded as handling coding. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
How much compute was used to train Composer 2?
Around 2.3 × 10²⁵ FLOP, on NVIDIA B300 (Blackwell Ultra). 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.
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.