You see how the bottom of the page says it’s made with GraphCast? Go to Google DeepMind’s page and read all about the GenCast and GraphCast foundation models, which are massive transformer-based and transformer-adjacent (GraphCast is a graph neural network using an encoder-decoder
December 18, 2025 at 2:42 PM
You see how the bottom of the page says it’s made with GraphCast? Go to Google DeepMind’s page and read all about the GenCast and GraphCast foundation models, which are massive transformer-based and transformer-adjacent (GraphCast is a graph neural network using an encoder-decoder
New feature for AI models (GraphCast & AIFS) on PolarWx - using the available pressure levels, these maps show the **estimated** potential precipitation type:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
February 5, 2025 at 4:10 AM
New feature for AI models (GraphCast & AIFS) on PolarWx - using the available pressure levels, these maps show the **estimated** potential precipitation type:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
A year ago it was GraphCast.
Today it's GenCast at Nature
www.nature.com/articles/s41...
Soon we'll see #AI powered weather forecasting at unprecedented level of accuracy....and that will pertain to other domains (like medicine)
@deep-mind.bsky.social
science.org/doi/10.1126/sc…
Today it's GenCast at Nature
www.nature.com/articles/s41...
Soon we'll see #AI powered weather forecasting at unprecedented level of accuracy....and that will pertain to other domains (like medicine)
@deep-mind.bsky.social
science.org/doi/10.1126/sc…
December 4, 2024 at 6:11 PM
A year ago it was GraphCast.
Today it's GenCast at Nature
www.nature.com/articles/s41...
Soon we'll see #AI powered weather forecasting at unprecedented level of accuracy....and that will pertain to other domains (like medicine)
@deep-mind.bsky.social
science.org/doi/10.1126/sc…
Today it's GenCast at Nature
www.nature.com/articles/s41...
Soon we'll see #AI powered weather forecasting at unprecedented level of accuracy....and that will pertain to other domains (like medicine)
@deep-mind.bsky.social
science.org/doi/10.1126/sc…
AI weather models have their strengths - but just like any other model, they also have their weaknesses.
This upcoming Saturday is a major failure for the AIFS & GraphCast - both consistently showed a major nor'easter (with little support from the GFS & ECMWF), only to completely lose the storm:
This upcoming Saturday is a major failure for the AIFS & GraphCast - both consistently showed a major nor'easter (with little support from the GFS & ECMWF), only to completely lose the storm:
February 26, 2025 at 12:12 AM
AI weather models have their strengths - but just like any other model, they also have their weaknesses.
This upcoming Saturday is a major failure for the AIFS & GraphCast - both consistently showed a major nor'easter (with little support from the GFS & ECMWF), only to completely lose the storm:
This upcoming Saturday is a major failure for the AIFS & GraphCast - both consistently showed a major nor'easter (with little support from the GFS & ECMWF), only to completely lose the storm:
We are still evaluating how the new suite of AI models may compare (e.g., AIFS, Google-DeepMind, Graphcast). Thus far, it looks like the same limitations that apply the ordinary NWP modeling also applies to the new suite of AI modeling.
August 10, 2025 at 3:44 PM
We are still evaluating how the new suite of AI models may compare (e.g., AIFS, Google-DeepMind, Graphcast). Thus far, it looks like the same limitations that apply the ordinary NWP modeling also applies to the new suite of AI modeling.
You may have seen a "GraphCast-GFS" model on some real-time websites - development on that model at NOAA has now led to the "AI-GFS" model which is scheduled to become operational on December 17th.
Output from the AI-GFS model is now available on PolarWx: polarwx.com/models/?mode...
Output from the AI-GFS model is now available on PolarWx: polarwx.com/models/?mode...
December 11, 2025 at 12:45 AM
You may have seen a "GraphCast-GFS" model on some real-time websites - development on that model at NOAA has now led to the "AI-GFS" model which is scheduled to become operational on December 17th.
Output from the AI-GFS model is now available on PolarWx: polarwx.com/models/?mode...
Output from the AI-GFS model is now available on PolarWx: polarwx.com/models/?mode...
Ein neues KI-Wettermodell von Google übertrifft selbst die Prognosen weltweit führender Institute. Forscher sprechen von einem "Wendepunkt der Wettervorhersage". https://www.zeit.de/digital/internet/2023-11/graphcast-wettervorhersage-kuenstliche-intelligenz-google
GraphCast: Diese KI trifft bessere Wettervorhersagen als ein Supercomputer
Ein neues KI-Wettermodell von Google übertrifft selbst die Prognosen weltweit führender Institute. Forscher sprechen von einem "Wendepunkt der Wettervorhersage".
www.zeit.de
November 26, 2023 at 1:23 PM
Ein neues KI-Wettermodell von Google übertrifft selbst die Prognosen weltweit führender Institute. Forscher sprechen von einem "Wendepunkt der Wettervorhersage". https://www.zeit.de/digital/internet/2023-11/graphcast-wettervorhersage-kuenstliche-intelligenz-google
Final set of additions to PolarWx for the time being - two AI emulator models (ECMWF's AIFS, and the GFS-based GraphCast) are now plotting in real time:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
January 21, 2025 at 1:35 AM
Final set of additions to PolarWx for the time being - two AI emulator models (ECMWF's AIFS, and the GFS-based GraphCast) are now plotting in real time:
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
polarwx.com/models/?mode...
Yes, an LLM a is a transformer model just like GenCast is, which has the same benefits as GraphCast but is used for wider-scale climate-level modeling because that architecture is better for that sort of scale.
GraphCast is transformer-adjacent. It doesn’t use self-attention for parallelization,
GraphCast is transformer-adjacent. It doesn’t use self-attention for parallelization,
December 18, 2025 at 2:56 PM
Yes, an LLM a is a transformer model just like GenCast is, which has the same benefits as GraphCast but is used for wider-scale climate-level modeling because that architecture is better for that sort of scale.
GraphCast is transformer-adjacent. It doesn’t use self-attention for parallelization,
GraphCast is transformer-adjacent. It doesn’t use self-attention for parallelization,
"Schneller, genauer, günstiger - so bewirbt Google sein Produkt Graphcast".
"Die größten Defizite sieht er, wenn es darum geht, Extreme vorherzusagen, die bisher noch nicht aufgetreten sind."
Graphcast nutzt Mustererkennung. Keine Muster (Daten), keine Erkenntnis.
www.n-tv.de/wissen/Macht...
"Die größten Defizite sieht er, wenn es darum geht, Extreme vorherzusagen, die bisher noch nicht aufgetreten sind."
Graphcast nutzt Mustererkennung. Keine Muster (Daten), keine Erkenntnis.
www.n-tv.de/wissen/Macht...
"Mit beispielloser Genauigkeit": Macht KI die Wetterdienste demnächst überflüssig?
Künstliche Intelligenz wird mittlerweile in fast allen Bereichen des Lebens genutzt - auch bei der Wettervorhersage. Private Angebote wie "Graphcast" von Google machen Behörden wie dem Deutschen Wette...
www.n-tv.de
May 7, 2024 at 6:53 PM
"Schneller, genauer, günstiger - so bewirbt Google sein Produkt Graphcast".
"Die größten Defizite sieht er, wenn es darum geht, Extreme vorherzusagen, die bisher noch nicht aufgetreten sind."
Graphcast nutzt Mustererkennung. Keine Muster (Daten), keine Erkenntnis.
www.n-tv.de/wissen/Macht...
"Die größten Defizite sieht er, wenn es darum geht, Extreme vorherzusagen, die bisher noch nicht aufgetreten sind."
Graphcast nutzt Mustererkennung. Keine Muster (Daten), keine Erkenntnis.
www.n-tv.de/wissen/Macht...
Like its predecessor (GraphCast), the weights & code of GenCast have been made publicly available: github.com/google-deepm...
We’re looking forward to seeing how the community builds on this!
We’re looking forward to seeing how the community builds on this!
GitHub - google-deepmind/graphcast
Contribute to google-deepmind/graphcast development by creating an account on GitHub.
github.com
December 10, 2024 at 11:00 AM
Like its predecessor (GraphCast), the weights & code of GenCast have been made publicly available: github.com/google-deepm...
We’re looking forward to seeing how the community builds on this!
We’re looking forward to seeing how the community builds on this!
Using AI emulators to further my point, the GFS-based GraphCast also completely lacks the major snow band the GFS is showing:
March 12, 2025 at 1:39 AM
Using AI emulators to further my point, the GFS-based GraphCast also completely lacks the major snow band the GFS is showing:
Verifikation des ECMWF bzgl. aktueller MachineLearning-Wettermodelle wie AIFS und GraphCast:
www.ecmwf.int/en/about/med...
Interessant ist v.a. diese Grafik:
1/
www.ecmwf.int/en/about/med...
Interessant ist v.a. diese Grafik:
1/
December 4, 2024 at 6:32 PM
Verifikation des ECMWF bzgl. aktueller MachineLearning-Wettermodelle wie AIFS und GraphCast:
www.ecmwf.int/en/about/med...
Interessant ist v.a. diese Grafik:
1/
www.ecmwf.int/en/about/med...
Interessant ist v.a. diese Grafik:
1/
I'm attending #AMS2025. If you're interested in AI for Weather prediction(particularly GraphCast/GenCast) or in tropical cyclones, send me a DM!
January 12, 2025 at 6:50 PM
I'm attending #AMS2025. If you're interested in AI for Weather prediction(particularly GraphCast/GenCast) or in tropical cyclones, send me a DM!
🆕 Nouveau modèle disponible sur Météociel :
AI-GFS (Global-EAGLE-Solo) expérimental 0.25° qui fait partie du projet EAGLE par la NOAA.
Ce modèle utilise le système GraphCast développé par Google réentrainé avec les données d'assimilation GDAS
➡️ meteociel.fr/r/AI-GFS
AI-GFS (Global-EAGLE-Solo) expérimental 0.25° qui fait partie du projet EAGLE par la NOAA.
Ce modèle utilise le système GraphCast développé par Google réentrainé avec les données d'assimilation GDAS
➡️ meteociel.fr/r/AI-GFS
November 23, 2025 at 7:15 PM
🆕 Nouveau modèle disponible sur Météociel :
AI-GFS (Global-EAGLE-Solo) expérimental 0.25° qui fait partie du projet EAGLE par la NOAA.
Ce modèle utilise le système GraphCast développé par Google réentrainé avec les données d'assimilation GDAS
➡️ meteociel.fr/r/AI-GFS
AI-GFS (Global-EAGLE-Solo) expérimental 0.25° qui fait partie du projet EAGLE par la NOAA.
Ce modèle utilise le système GraphCast développé par Google réentrainé avec les données d'assimilation GDAS
➡️ meteociel.fr/r/AI-GFS
New video up on GraphCast, the weather forecasting model from Deepmind. www.youtube.com/watch?v=asMw...
Learning Better Weather Forecasts with AI
YouTube video by 5 Minute Papers on AI for the Planet
www.youtube.com
June 28, 2025 at 3:44 PM
New video up on GraphCast, the weather forecasting model from Deepmind. www.youtube.com/watch?v=asMw...
It's so annoying that people (even within DeepMind) keep saying that OpenAI is ahead "in AI", when DeepMind has models like AlphaFold, AlphaProof, GenCast and GraphCast, FunSearch, etc
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
September 21, 2024 at 7:15 PM
It's so annoying that people (even within DeepMind) keep saying that OpenAI is ahead "in AI", when DeepMind has models like AlphaFold, AlphaProof, GenCast and GraphCast, FunSearch, etc
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
AI for Science has such a clearer potential for helping people, but VCs and the media are obsessed by GenAI
I'll be at #AMS2025 next week alongside some other Google DeepMind colleagues behind GenCast/GraphCast. Excited to meet people, discuss ML for weather, and learn!
January 10, 2025 at 4:24 PM
I'll be at #AMS2025 next week alongside some other Google DeepMind colleagues behind GenCast/GraphCast. Excited to meet people, discuss ML for weather, and learn!
ML weather models like GraphCast & NeuralGCM show promise in weather forecasting, but can they be used for data assimilation? Our study looks into their tangent linear & adjoint models, finding challenges in robustness & unphysical noise. 🌦️ #AI #NWP
arxiv.org/abs/2411.14677
arxiv.org/abs/2411.14677
Exploring the Use of Machine Learning Weather Models in Data Assimilation
The use of machine learning (ML) models in meteorology has attracted significant attention for their potential to improve weather forecasting efficiency and accuracy. GraphCast and NeuralGCM, two prom...
arxiv.org
November 25, 2024 at 2:23 AM
ML weather models like GraphCast & NeuralGCM show promise in weather forecasting, but can they be used for data assimilation? Our study looks into their tangent linear & adjoint models, finding challenges in robustness & unphysical noise. 🌦️ #AI #NWP
arxiv.org/abs/2411.14677
arxiv.org/abs/2411.14677
Yeah I think I was confusing GraphCast (the older model this work is based on) with GenCast which is deepmind's latest work which is transformer and diffusion based. Won't be long before that's integrated too though. It's designed to be pretty modular
December 18, 2025 at 4:42 PM
Yeah I think I was confusing GraphCast (the older model this work is based on) with GenCast which is deepmind's latest work which is transformer and diffusion based. Won't be long before that's integrated too though. It's designed to be pretty modular
Interesting discrepancy in the NWP guidance for this Wed's potential brief bout of freezing rain in the OKC area. Most of the models are very marginal or slightly above freezing, but the AI-based models I've seen (GFS GraphCast, AIFS) are significantly colder than their physics-based counterparts.
February 10, 2025 at 9:39 PM
Interesting discrepancy in the NWP guidance for this Wed's potential brief bout of freezing rain in the OKC area. Most of the models are very marginal or slightly above freezing, but the AI-based models I've seen (GFS GraphCast, AIFS) are significantly colder than their physics-based counterparts.
Inspired by Google Deepmind's weather forecasting model GraphCast, we also developed an organelle graph autoencoder using Nellie's outputs to analyze mitochondrial networks' responses to ionomycin treatment, revealing complex dynamics invisible to traditional analysis
🧵4/N
🧵4/N
February 27, 2025 at 8:13 PM
Inspired by Google Deepmind's weather forecasting model GraphCast, we also developed an organelle graph autoencoder using Nellie's outputs to analyze mitochondrial networks' responses to ionomycin treatment, revealing complex dynamics invisible to traditional analysis
🧵4/N
🧵4/N
This Is promising! “GraphCast” can make a 10-day forecast in less than a minute; traditional models take hours on a supercomputers. GraphCast was about 10 percent more accurate than the European model on more than 90 percent of the weather variables evaluated. www.washingtonpost.com/weather/2023...
Why your weather forecasts may soon become more accurate
Google has produced a weather forecasting model using artificial intelligence with better accuracy, faster speed and lower costs, study finds.
www.washingtonpost.com
November 14, 2023 at 6:10 PM
This Is promising! “GraphCast” can make a 10-day forecast in less than a minute; traditional models take hours on a supercomputers. GraphCast was about 10 percent more accurate than the European model on more than 90 percent of the weather variables evaluated. www.washingtonpost.com/weather/2023...
わずか1分で10日間の天気を予測可能なAI「GraphCast」をGoogle DeepMindが発表、スパコンで数時間かけた予測より高精度
https://gigazine.net/news/20231115-google-graphcast-global-weather-forecasting/
https://gigazine.net/news/20231115-google-graphcast-global-weather-forecasting/
わずか1分で10日間の天気を予測可能なAI「GraphCast」をGoogle DeepMindが発表、スパコンで数時間かけた予測より高精度
Google DeepMindが気象予測AI「GraphCast」を発表しました。10日間の天気を予測する場合、従来の気象予測システムではスーパーコンピューターを数時間稼働させる必要がありましたが、GraphCastでは1台のマシンを1分稼働させるだけで従来の予測システムを超える精度での予測が可能とアピールされています。
gigazine.net
November 15, 2023 at 2:09 AM
わずか1分で10日間の天気を予測可能なAI「GraphCast」をGoogle DeepMindが発表、スパコンで数時間かけた予測より高精度
https://gigazine.net/news/20231115-google-graphcast-global-weather-forecasting/
https://gigazine.net/news/20231115-google-graphcast-global-weather-forecasting/