GNMT
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
- Organisation type
- Industry
- Country
- United States of America
- Published
- 26 September 2016
- Authors
- Yonghui Wu, Mike Schuster, Zhifeng Chen, Quoc V. Le, Mohammad Norouzi, Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey, Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, Łukasz Kaiser, Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens, George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa, Alex Rudnick, Oriol Vinyals, Greg Cor…
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Translation
- Approach
- Reinforcement learning
- Numerical format
- FP32
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
- 278M
- Training data
- 720,000,000 tokens
- Epochs
- 1
Table 5 in 'Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer' https://arxiv.org/abs/1701.06538
[WORDS] " On WMT En→Fr, the training set contains 36M sentence pairs. On WMT En→De, the training set contains 5M sentence pairs." "we also test GNMT on Google’s translation production corpora, which are two to three decimal orders of magnitudes bigger than the WMT corpora for a given language pair." 41M sentence pairs * 2 sentences per pair * 15 words/sentence * 10^2.5
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
- 6.6 × 10²¹ FLOP
- How it was established
- Hardware,Third-party estimation
From AI and Compute: "sqrt(10 * 100) factor added because production model used 2-3 orders of magnitude more data, but only 1 epoch rather than 10. 96 K80 GPU’s * 9 days * 8.5 TFLOPS * 0.33 utilization * sqrt(10 * 100) = 6.9e6 PF = 79 pfs-days" source: https://openai.com/blog/ai-and-compute/ https://www.wolframalpha.com/input?i=96+*+9+days+*+8.5+TFLOPS+*+0.33+*+sqrt%281000%29
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 Tesla K80
- Chip-hours
- 655,730
- Compute cost
- $201,332
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
- Hosted access (no API)
- Training code
- Unreleased
presumably deployed via Google translate
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Frontier model
- Yes
- Why it is tracked
- Highly cited
- Record confidence
- Likely
- Citations
- 7,254
Sources
Where this record came from and when it was last checked.
- Reference
- Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
- Last updated
- 25 May 2026
What the numbers mean
Where it came from
GNMT was published by Google, in the country recorded as United States of America, during September 2016. The category the publisher falls under is industry.
It works in the domain of Language, and is recorded as performing the task of translation.
Its weights were never published, so it can only be reached through its provider. No graphics card changes that.
How it was trained
The training run consumed about 6.6 × 10²¹ FLOP, on hardware recorded as NVIDIA Tesla K80. That figure measures what producing the model cost, and has no bearing on how fast it answers.
The training set ran to roughly 720,000,000 tokens of text.
Its inclusion criterion: highly cited.
Answers
GNMT — common questions
GNMT— how much compute was used to train it?
Training consumed around 6.6 × 10²¹ FLOP, on hardware recorded as NVIDIA Tesla K80. 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.
GNMT— what GPU do I need to run it?
None. This 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.
GNMT— is it open source?
No. Its weights have not been published, so it exists only as a service controlled by its owner.
GNMT— how many parameters does it have?
It has a parameter count of 278M. Table 5 in 'Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer' https://arxiv.org/abs/1701.06538. 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.
GNMT— who created it?
It was published by Google, based in United States of America, an organisation categorised as industry.
GNMT— when was it released?
It was published in September 2016. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
GNMT— what is it used for?
It works in the domain of Language, and is recorded as handling the task of translation. 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.