Identity
- Type
- frontier lab
- Country label
- UK / USA
- HQ country
- UK
- Founding country
- UK
- Parent
Google- Parent country
- USA
- Country basis
- the research notes
Where this comes from
Convention D1: a dual label, HQ first, when the HQ, founding or parent country differ.
Models by Class
Models
ModelClassOpennessParametersReleased
Nano Banana 2.1Google DeepMind, Googleimage generation6 Oct 2026Gemini 4 ArgonLLM30 Sep 2026Gemini 3.8 FlashLLM2 Sep 2026Gemini 3.7 FlashLLMAug 2026WeatherNext 3Google DeepMind, GoogleweatherAug 2026Gemini Robotics 2robot policy30 Jul 2026Gemini Robotics 2 VLArobot policy30 Jul 2026Gemini Robotics-ER 2multimodal LLM30 Jul 2026Gemini Robotics On-Device 2robot policy30 Jul 2026Gemini 3.6 FlashLLM21 Jul 2026Magenta RealTime 2music and audioJun 2026Gemini Omni FlashGoogle DeepMind, Googlevideo generation19 May 2026Gemini 3.5 Flashmultimodal LLMMay 2026Gemini Robotics-ER 1.6Google DeepMind, Boston Dynamicsmultimodal LLM14 Apr 2026Gemma 4Google, Google DeepMindLLM2 Apr 2026Gemma 4 31BGoogle, Google DeepMindmultimodal LLM2 Apr 2026Gemma 4 E2BGoogle, Google DeepMindedge LLM2 Apr 2026Gemma 4 E4BGoogle, Google DeepMindedge LLM2 Apr 2026Veo 3.1 Litevideo generationMar 2026Gemini 3.1 Pro (preview)Google DeepMind, Googlemultimodal LLM19 Feb 2026Project GenieGoogle DeepMind, Googleworld model29 Jan 2026Veo 3.1video generationJan 2026Gemini 3 Deep Thinkreasoning18 Nov 2025WeatherNext 2 (FGN)Google DeepMind, Googleweather17 Nov 2025Natural Forests of the World 2020 model (MTSViT)Google DeepMind, Google, World Resources Institute (WRI), IIASAgeospatial13 Nov 2025
C2S-Scale-Gemma-2-2BYale University, Google, Google DeepMindscientific6 Oct 2025C2S-Scale-Gemma-2-27BYale University, Google, Google DeepMindscientific6 Oct 2025Gemini Robotics 1.5robot policy25 Sep 2025Gemini Robotics-ER 1.5multimodal LLM25 Sep 2025Dreamer 4world modelSep 2025Genie 3world model5 Aug 2025AlphaEarth Foundations (Satellite Embedding v1 Annual)geospatialJul 2025Gemini Robotics On-Devicerobot policy24 Jun 2025Gemma 3n E2Bedge LLM12 Jun 2025Gemma 3n E4Bedge LLM3 Jun 2025DreamerV3world model2 Apr 2025Gemini Roboticsrobot policy12 Mar 2025Gemini Robotics-ERmultimodal LLM12 Mar 2025Gemini 3 ProGoogle DeepMind, Googlemultimodal LLM2025Veo 2video generation16 Dec 2024Genie 2world modelDec 2024WeatherNext 1 Gen (GenCast)Google DeepMind, GoogleweatherDec 2024
Imagen 3image generation14 May 2024AlphaFold 3Google DeepMind, Isomorphic Labsscientific8 May 2024Genieworld modelMar 2024GenCast (WeatherNext 1 Gen)weather2024WeatherNext 1 Graph (GraphCast)Google DeepMind, GoogleweatherNov 2023NaViT and FlexiViT (NaFlexViT)vision12 Jul 2023AlphaDevreinforcement learning7 Jun 2023GraphCast (WeatherNext 1 Graph)weather2023RT-2-Xrobot policy2023AlphaTensorreinforcement learning5 Oct 2022Perceiver (vision)vision30 Jul 2021NFNetvision11 Feb 2021AlphaFold 2scientificNov 2020BYOLvision13 Jun 2020Agent57reinforcement learning30 Mar 2020MuZeroreinforcement learning19 Nov 2019BigBiGANimage generation4 Jul 2019AlphaStarreinforcement learning24 Jan 2019BigGAN (incl. BigGAN-Deep)image generation28 Sep 2018SPIRALimage generation3 Apr 2018WaveRNNspeech synthesis23 Feb 2018IMPALAreinforcement learning5 Feb 2018AlphaZeroreinforcement learning5 Dec 2017VQ-VAEimage generation2 Nov 2017I3D (Inflated 3D ConvNet)vision22 May 2017WaveNetspeech synthesis12 Sep 2016AlphaGoreinforcement learningOct 2015DQN (Deep Q-Network)reinforcement learning19 Dec 2013AMIEGoogle, Google DeepMindmultimodal LLM–AlphaGenomescientific–antigravity-preview-09-2026LLM–deep-research-max-preview-04-2026LLM–deep-research-preview-04-2026LLM–Dreamerworld model–GeminiGoogle DeepMind, Googlemultimodal LLM–Gemini 2.0multimodal LLM–gemini-2.5-computer-use-preview-10-2025LLM–Gemini 2.5 FlashGoogle DeepMind, Googlemultimodal LLM–Gemini 2.5 Flash-LiteLLM–Gemini 2.5 ProLLM–Gemini 3.1 Flash-LiteLLM–Gemini 3.5 ProLLM–gemini-3.8-livespeech synthesis–gemini-3.8-live-extended-thinkingspeech synthesis–gemini-robotics-er-2-previewmultimodal LLM–Gemini Roboticsrobot policy–Genieworld model–
Imagenimage generation–Imagen 4image generation–Lyriamusic and audio–Lyria 2music and audio–Lyria 3music and audio–Lyria 3.5music and audio–Lyria 3 Promusic and audio–Lyriamusic and audio–Lyria RealTimemusic and audio–RT-2robot policy–Veovideo generation–Veo 3video generation–Veo 3.1 Fastvideo generation–Gemma 3nedge LLM–Gemini 3.5 Flash-Litemultimodal LLM–Gemini 3.8 Flash CyberLLM–C2S-Scale (Cell2Sentence)Yale University, Google, Google DeepMindscientific–Links
ParentGooglebig tech, USA
Googlebig tech, USAOrganisationGoogle DeepMindfrontier lab, UK / USA
Google DeepMindfrontier lab, UK / USAModelsNano Banana 2.1Google DeepMind, image generationGemini 4 ArgonGoogle DeepMind, LLMGemini 3.8 FlashGoogle DeepMind, LLMGemini 3.7 FlashGoogle DeepMind, LLM+102 more
Nano Banana 2.1Google DeepMind, image generationGemini 4 ArgonGoogle DeepMind, LLMGemini 3.8 FlashGoogle DeepMind, LLMGemini 3.7 FlashGoogle DeepMind, LLM+102 more
Models 106
GeminiGoogle DeepMind, Google, multimodal LLMDeveloperGemini RoboticsGoogle DeepMind, robot policyDeveloperGemma 4Google, Google DeepMind, LLMDeveloperAMIEGoogle, Google DeepMind, multimodal LLMDeveloperGemini Robotics-ER 1.6Google DeepMind, Boston Dynamics, multimodal LLMDeveloperGemma 4 E2BGoogle, Google DeepMind, edge LLMDeveloperC2S-Scale (Cell2Sentence)Yale University, Google, Google DeepMind, scientificDeveloperC2S-Scale-Gemma-2-2BYale University, Google, Google DeepMind, scientificDeveloperGemini 3.6 FlashGoogle DeepMind, LLMDeveloperGemma 3nGoogle DeepMind, edge LLMDeveloperGenie 3Google DeepMind, world modelDeveloperLyriaGoogle DeepMind, music and audioDeveloper+94
Nano Banana 2.1Google DeepMind, Google, image generationDeveloperNatural Forests of the World 2020 model (MTSViT)Google DeepMind, Google, World Resources Institute (WRI), IIASA, geospatialDeveloperWeatherNext 2 (FGN)Google DeepMind, Google, weatherDeveloperWeatherNext 3Google DeepMind, Google, weatherDeveloperC2S-Scale-Gemma-2-27BYale University, Google, Google DeepMind, scientificDeveloperGemini 3 ProGoogle DeepMind, Google, multimodal LLMDeveloperGemini 3.8 FlashGoogle DeepMind, LLMDeveloperGemini RoboticsGoogle DeepMind, robot policyDeveloperGemini Robotics 2Google DeepMind, robot policyDeveloperGemma 3n E2BGoogle DeepMind, edge LLMDeveloperGemma 3n E4BGoogle DeepMind, edge LLMDeveloperLyria 3Google DeepMind, music and audioDeveloperVeo 3.1Google DeepMind, video generationDeveloperAlphaFold 3Google DeepMind, Isomorphic Labs, scientificDeveloperGemini 2.5 FlashGoogle DeepMind, Google, multimodal LLMDeveloperGemini 3.1 Pro (preview)Google DeepMind, Google, multimodal LLMDeveloperGemini 3.5 FlashGoogle DeepMind, multimodal LLMDeveloperGemini 3.7 FlashGoogle DeepMind, LLMDeveloperGemini Omni FlashGoogle DeepMind, Google, video generationDeveloperGemini Robotics 1.5Google DeepMind, robot policyDeveloperGemini Robotics-ERGoogle DeepMind, multimodal LLMDeveloperGemini Robotics-ER 1.5Google DeepMind, multimodal LLMDeveloperGemini Robotics-ER 2Google DeepMind, multimodal LLMDevelopergemini-2.5-computer-use-preview-10-2025Google DeepMind, LLMDeveloperGemma 4 31BGoogle, Google DeepMind, multimodal LLMDeveloperGemma 4 E4BGoogle, Google DeepMind, edge LLMDeveloperGenieGoogle DeepMind, world modelDeveloperGenie 2Google DeepMind, world modelDeveloperLyria 2Google DeepMind, music and audioDeveloperVeoGoogle DeepMind, video generationDeveloperAlphaZeroGoogle DeepMind, reinforcement learningDeveloperDreamerGoogle DeepMind, world modelDeveloperDreamer 4Google DeepMind, world modelDeveloperDreamerV3Google DeepMind, world modelDeveloperGemini 2.5 ProGoogle DeepMind, LLMDeveloperGemini 3.8 Flash CyberGoogle DeepMind, LLMDeveloperGemini 4 ArgonGoogle DeepMind, LLMDeveloperGemini Robotics 2 VLAGoogle DeepMind, robot policyDeveloperGemini Robotics On-DeviceGoogle DeepMind, robot policyDeveloperGemini Robotics On-Device 2Google DeepMind, robot policyDeveloperGenieGoogle DeepMind, world modelDeveloperImagenGoogle DeepMind, image generationDeveloperImagen 4Google DeepMind, image generationDeveloperLyriaGoogle DeepMind, music and audioDeveloperLyria 3.5Google DeepMind, music and audioDeveloperMuZeroGoogle DeepMind, reinforcement learningDeveloperProject GenieGoogle DeepMind, Google, world modelDeveloperVeo 3Google DeepMind, video generationDeveloperWeatherNext 1 Gen (GenCast)Google DeepMind, Google, weatherDeveloperWeatherNext 1 Graph (GraphCast)Google DeepMind, Google, weatherDeveloperAlphaEarth Foundations (Satellite Embedding v1 Annual)Google DeepMind, geospatialDeveloperAlphaFold 2Google DeepMind, scientificDeveloperAlphaGenomeGoogle DeepMind, scientificDeveloperantigravity-preview-09-2026Google DeepMind, LLMDeveloperBigBiGANGoogle DeepMind, image generationDeveloperdeep-research-max-preview-04-2026Google DeepMind, LLMDeveloperdeep-research-preview-04-2026Google DeepMind, LLMDeveloperGemini 2.0Google DeepMind, multimodal LLMDeveloperGemini 2.5 Flash-LiteGoogle DeepMind, LLMDeveloperGemini 3 Deep ThinkGoogle DeepMind, reasoningDeveloperGemini 3.1 Flash-LiteGoogle DeepMind, LLMDeveloperGemini 3.5 Flash-LiteGoogle DeepMind, multimodal LLMDeveloperGemini 3.5 ProGoogle DeepMind, LLMDevelopergemini-3.8-liveGoogle DeepMind, speech synthesisDevelopergemini-3.8-live-extended-thinkingGoogle DeepMind, speech synthesisDeveloperGenCast (WeatherNext 1 Gen)Google DeepMind, weatherDeveloperGraphCast (WeatherNext 1 Graph)Google DeepMind, weatherDeveloperImagen 3Google DeepMind, image generationDeveloperLyria 3 ProGoogle DeepMind, music and audioDeveloperLyria RealTimeGoogle DeepMind, music and audioDeveloperRT-2Google DeepMind, robot policyDeveloperRT-2-XGoogle DeepMind, robot policyDeveloperVeo 2Google DeepMind, video generationDeveloperVeo 3.1 FastGoogle DeepMind, video generationDeveloperVeo 3.1 LiteGoogle DeepMind, video generationDeveloperAgent57Google DeepMind, reinforcement learningDeveloperAlphaDevGoogle DeepMind, reinforcement learningDeveloperAlphaGoGoogle DeepMind, reinforcement learningDeveloperAlphaStarGoogle DeepMind, reinforcement learningDeveloperAlphaTensorGoogle DeepMind, reinforcement learningDeveloperBigGAN (incl. BigGAN-Deep)Google DeepMind, image generationDeveloperBYOLGoogle DeepMind, visionDeveloperDQN (Deep Q-Network)Google DeepMind, reinforcement learningDevelopergemini-robotics-er-2-previewGoogle DeepMind, multimodal LLMDeveloperI3D (Inflated 3D ConvNet)Google DeepMind, visionDeveloperIMPALAGoogle DeepMind, reinforcement learningDeveloperMagenta RealTime 2Google DeepMind, music and audioDeveloperNaViT and FlexiViT (NaFlexViT)Google DeepMind, visionDeveloperNFNetGoogle DeepMind, visionDeveloperPerceiver (vision)Google DeepMind, visionDeveloperSPIRALGoogle DeepMind, image generationDeveloperVQ-VAEGoogle DeepMind, image generationDeveloperWaveNetGoogle DeepMind, speech synthesisDeveloperWaveRNNGoogle DeepMind, speech synthesisDeveloperOrganisations 1
Scores
| Benchmark | Model | Score | Who ran it | Setting |
|---|---|---|---|---|
| ARC-AGI-2 | Gemini 3 Deep Think | 45.1 % | vendor-reported | With code execution; ARC Prize Verified per vendor |
| ARC-AGI-2 | Gemini 3 Deep Think | 84.6 % | vendor-reported | Updated Deep Think; verified by the ARC Prize Foundation per vendor |
| ARC-c | Gemma 3n E2B | 51.7 % | vendor-reported | PT (pretrained checkpoint) |
| ARC-c | Gemma 3n E4B | 61.6 % | vendor-reported | PT (pretrained checkpoint) |
| Arena Text leaderboard | Gemini 3.1 Pro (preview) | 1,487 Elo | independent | preview |
| Arena Text leaderboard | Gemini 3.8 Flash | 1,497 Elo | independent | high |
| Arena Text leaderboard | Gemini 4 Argon | 1,525 Elo | independent | high |
| Arena Vision leaderboard | Gemini 3.7 Flash | 1,297 Elo | independent | high; preliminary |
| Arena Vision leaderboard | Gemini 3.8 Flash | 1,289 Elo | independent | high; preliminary |
| Artificial Analysis Intelligence Index | Gemini 3.1 Pro (preview) | 57 index | independent | v4.0, stale; superseded by AA v4.3.2 |
| Artificial Analysis Intelligence Index | Gemini 3.8 Flash | 41 index | independent | high; 40.93 |
| Artificial Analysis Intelligence Index | Gemini 4 Argon | 53 index | independent | high; 52.56 per secondary |
| Artificial Analysis Music Arena | Lyria 2 | 932 Elo | independent | secondary capture of AA post |
| Artificial Analysis Music Arena | Lyria 3 Pro | 1,078 Elo | independent | secondary capture of AA post; tied 3-6 |
| Artificial Analysis Music Arena | Lyria 3.5 | 1,076 Elo | independent | secondary capture of AA post; tied 3-6 |
| Artificial Analysis Text-to-Image Arena | Nano Banana 2.1 | 1,160 Elo | independent | rank shown as 4-6 |
| Artificial Analysis Text-to-Video leaderboard | Veo 3.1 | 962 Elo | independent | |
| AutomationBench | Gemini 4 Argon | 51.3 % | vendor-reported | Google self-reported |
| benchlm.ai frontier composite | Gemini 4 Argon | 81.8 index | vendor-reported | estimated |
| Codeforces rating | Gemini 3 Deep Think | 3,455 Elo | vendor-reported | Updated Deep Think |
| DeepSWE | Gemini 3.8 Flash | 70 % | vendor-reported | above 70% (chart only) |
| DeepSWE | Gemini 4 Argon | 77.9 % | vendor-reported | Google self-reported |
| DeepSWE (version not stated) | Gemini 3.5 Flash | 37 % | vendor-reported | |
| DeepSWE (version not stated) | Gemini 3.6 Flash | 49 % | vendor-reported | |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics 2 VLA | 45.7 % | vendor-reported | Apollo 2 whole-body pick, floor |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics 2 VLA | 68.4 % | vendor-reported | Apollo 2 whole-body pick, table |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics 2 VLA | 74.2 % | vendor-reported | Franka Duo pick-and-place |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics 2 VLA | 76.3 % | vendor-reported | Apollo 2 whole-body pick, shelf |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics 2 VLA | 89.6 % | vendor-reported | Franka Duo precise insertion |
| Gemini Robotics 2 vendor task evaluations | Gemini Robotics-ER 2 | 91.3 % | vendor-reported | Moment finding |
| Global MMLU | Gemma 3n E2B | 55.1 % | vendor-reported | IT |
| Global MMLU | Gemma 3n E4B | 60.3 % | vendor-reported | IT |
| GPQA | Gemma 4 31B | 84.3 % | vendor-reported | Reported as 'GPQA' by secondary source; subset not stated |
| GPQA Diamond | Gemini 3 Deep Think | 93.8 % | vendor-reported | |
| GPQA Diamond | Gemma 3n E2B | 24.8 % | vendor-reported | IT |
| GPQA Diamond | Gemma 3n E4B | 23.7 % | vendor-reported | IT |
| GPQA Diamond | Gemma 4 31B | 84.3 % | vendor-reported | |
| Harvey Legal Agent Benchmark | Gemini 4 Argon | 19.6 | vendor-reported | Google self-reported |
| HellaSwag | Gemma 3n E2B | 72.2 % | vendor-reported | PT (pretrained checkpoint) |
| HellaSwag | Gemma 3n E4B | 78.6 % | vendor-reported | PT (pretrained checkpoint) |
| Humanity's Last Exam (HLE) | Gemini 3 Deep Think | 41 % | vendor-reported | Without tools |
| Humanity's Last Exam (HLE) | Gemini 3 Deep Think | 48.4 % | vendor-reported | Without tools; updated Deep Think |
| LABBench 2 | Gemini 4 Argon | 88.8 | vendor-reported | Google self-reported |
| LiveCodeBench v5 | Gemma 3n E2B | 18.6 % | vendor-reported | IT |
| LiveCodeBench v5 | Gemma 3n E4B | 25.7 % | vendor-reported | IT |
| LMArena Text-to-Image Arena | Nano Banana 2.1 | 1,328 Elo | independent | preliminary |
| LMArena Text-to-Video Arena | Gemini Omni Flash | 1,513 Elo | independent | |
| LMArena Text-to-Video Arena | Veo 3.1 | 1,364 Elo | independent | veo-3.1-audio |
| MGSM | Gemma 3n E2B | 53.1 % | vendor-reported | IT |
| MGSM | Gemma 3n E4B | 60.7 % | vendor-reported | IT |
| MMLU-Pro | Gemma 4 31B | 85.2 % | vendor-reported | |
| MMMU | Gemini 3 Pro | 87.2 % | vendor-reported | |
| OmniDocBench | Gemini 3 Pro | 92.91 score | vendor-reported | overall |
| PoseBusters | AlphaFold 3 | rank only | vendor-reported | Beats Vina docking (P = 2.27e-13); success rate shown only in a figure |
| Real5-OmniDocBench | Gemini 3 Pro | 89.24 score | vendor-reported | overall |
| SEA-HELM | Gemma 4 31B | 76.17 score | maintainer-run | |
| SEA-HELM | Gemma 4 31B | 79.75 index | maintainer-run | Indonesian |
| SEA-HELM | Gemma 4 31B | 79.75 score | maintainer-run | |
| SEA-HELM | Gemma 4 E2B | 58.82 score | maintainer-run | |
| SEA-HELM | Gemma 4 E2B | 64.13 score | maintainer-run | |
| SEA-HELM | Gemma 4 E4B | 66.3 score | maintainer-run | Checkpoint gemma-4-E4B-it as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps |
| SEA-HELM | Gemma 4 E4B | 69.65 score | maintainer-run | Checkpoint gemma-4-E4B-it as listed; release view (data last updated 18 Sep 2026); score is the average of 2000 bootstraps |
| SWE-bench Pro | Gemini 3.1 Pro (preview) | 46.1 % | maintainer-run | Scale SEAL public, mini-swe-agent |
| SWE-bench Verified | Gemini 3.1 Pro (preview) | 80.6 % | vendor-reported | |
| Vals Index | Gemini 4 Argon | 68.9 | vendor-reported | Google self-reported |
| VBench | Veo 3 | 85.06 % | maintainer-run | Total score; submission-based board |
| VBench | Veo 3 | 85.7 % | maintainer-run | Quality score; submission-based board |
| VBench-2.0 | Veo 3 | 60.85 % | maintainer-run | Creativity score |
| VBench-2.0 | Veo 3 | 66.72 % | maintainer-run | Total score |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 1.477 RMSE K | maintainer-run | T850 day 5, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 123.5 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, GenCast (mean) vs ERA5 |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 124.3 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 2.693 RMSE K | maintainer-run | T850 day 10, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 262.4 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, GenCast (mean) vs ERA5 |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 264.7 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 39.2 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 39.3 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, GenCast (mean) vs ERA5 |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 606.5 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, GenCast (mean) vs ERA5 |
| WeatherBench 2 | GenCast (WeatherNext 1 Gen) | 607.5 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, GenCast (oper.) (mean) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 1.555 RMSE K | maintainer-run | T850 day 5, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 120.6 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 124.2 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, GraphCast vs ERA5 |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 270.2 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 274.3 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, GraphCast vs ERA5 |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 3.364 RMSE K | maintainer-run | T850 day 10, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 38.3 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 39.8 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, GraphCast vs ERA5 |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 731.9 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, GraphCast vs ERA5 |
| WeatherBench 2 | GraphCast (WeatherNext 1 Graph) | 747.6 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, GraphCast (oper.) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 1.348 RMSE K | maintainer-run | T850 day 5, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 1.81 RMSE K | maintainer-run | T850 day 5, deterministic, FGN (oper.) (1st member) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 104.5 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 139.4 RMSE m2/s2 | maintainer-run | Z500 day 3, deterministic, FGN (oper.) (1st member) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 2.595 RMSE K | maintainer-run | T850 day 10, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 233.1 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 3.632 RMSE K | maintainer-run | T850 day 10, deterministic, FGN (oper.) (1st member) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 313.7 RMSE m2/s2 | maintainer-run | Z500 day 5, deterministic, FGN (oper.) (1st member) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 32.3 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 43.7 RMSE m2/s2 | maintainer-run | Z500 day 1, deterministic, FGN (oper.) (1st member) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 583.1 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, FGN (oper.) (mean) vs Analysis |
| WeatherBench 2 | WeatherNext 2 (FGN) | 811.7 RMSE m2/s2 | maintainer-run | Z500 day 10, deterministic, FGN (oper.) (1st member) vs Analysis |
Where this comes from
Every recorded score of this organisation's models, with who ran it. Scores from different settings are not directly comparable.
Sources
- Models | Gemini API | Google AI for Developers ai.google.dev https://ai.google.dev/gemini-api/docs/models
- LLM Leaderboard - Comparison of AI models from OpenAI, Anthropic, Google, SpaceXAI & others | Artificial Analysis artificialanalysis.ai https://artificialanalysis.ai/leaderboards/models
- WeatherNext 2: Our most advanced weather forecasting model blog.google https://blog.google/technology/google-deepmind/weathernext-2/
- GitHub - google-deepmind/alphafold3: AlphaFold 3 inference pipeline. github.com https://github.com/google-deepmind/alphafold3
- Google DeepMind - Wikipedia en.wikipedia.org https://en.wikipedia.org/wiki/Google_DeepMind
- Google Releases Gemma 4: Open Models Under Apache 2.0 datanorth.ai https://datanorth.ai/news/google-releases-gemma-4-open-models
- Gemma 3n , Google DeepMind deepmind.google https://deepmind.google/models/gemma/gemma-3n/
- Accurate structure prediction of biomolecular interactions with AlphaFold 3 - Nature nature.com https://www.nature.com/articles/s41586-024-07487-w
- WeatherNext 3 | Google for Developers developers.google.com https://developers.google.com/weathernext/guides/models
- Training compute of frontier AI models grows by 4-5x per year epoch.ai https://epoch.ai/blog/training-compute-of-frontier-ai-models-grows-by-4-5x-per-year
- Over 30 AI models have been trained at the scale of GPT-4 epoch.ai https://epoch.ai/data-insights/models-over-1e25-flop
- Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2 arxiv.org https://arxiv.org/abs/2408.05147
- DeepMind thinks its new Genie 3 world model presents a stepping stone toward AGI | TechCrunch techcrunch.com https://techcrunch.com/2025/08/05/deepmind-reveals-genie-3-a-world-model-that-could-be-the-key-to-reaching-agi
- deepmindsafetyresearch.medium.com/negative-results-for-sparse-autoencoders-on-d... https://deepmindsafetyresearch.medium.com/negative-results-for-sparse-autoencoders-on-downstream-tasks-and-deprioritising-sae-research-6cadcfc125b9
To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Google DeepMind", https://aurion.semai.id/organisations/google-deepmind/, records of October 2026.