Tech Meridian← ALL MODELS
PROMY MERIDIAN

GOOGLE / GOOGLE DEEPMIND · MODEL RELEASE TRACKER

Gemini 4 Argon

Gemini 4 Argon is a Gemini 4 frontier model from Google described as optimized for deep, long‑horizon reasoning and for software‑engineering and knowledge‑work workflows. It was released prior to the announcement of Gemini 4 Carbon and is referenced in Google’s Gemini 4 family materials.

CURRENT SNAPSHOT5/5 DIMENSIONS WITH DATA

The dimensions that change the decision.

PRICING
Introductory price (reported)Reported introductory price: $2 per million input tokens and $10 per million output tokens (stated as the same introductory price as Gemini 4 Argon).
CONTEXT WINDOW
Context window / token capabilityArgon is described as built with an industry‑leading one‑million‑token context capability (the reporting also states support for up to one million output tokens).
MODALITIES
ModalityDescribed as a multimodal physical-reasoning model combining visual inputs, language, and action/simulation for robotics (video + language + control).
BENCHMARKS
Benchmark reportingMedia reporting cites Google claims that Argon scored higher than competing frontier models (including OpenAI’s GPT‑6 Astra and Anthropic’s Fable/Opus) on a variety of AI benchmarks and is listed as leading on at least one model index cited by Google.
Benchmark (ERQA)Reported score of 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning without helper tools; reported by Google as leading other frontier models on that test.
Vals Index rankingGoogle reports Argon ranks #1 on the Vals Index, an internal measure mentioned as assessing model performance across economic sectors such as finance, coding, legal, and tax work.
Cybersecurity evaluationGoogle reports Argon solved all 12 end‑to‑end simulated cyberattack scenarios in their cybersecurity tests (up from five in the previous generation), and describes frontier‑level cyberdefense capabilities intended to help defenders.
RE‑Bench performanceOn DeepMind's RE‑Bench (74 realistic ML engineering tasks), Gemini 4 Argon solves about half of the tasks (~50%). By comparison, Gemini 3.7 Flash solved 27% on the same benchmark.
Embodied Reasoning (ERQA) scoreArgon scored 87% on the Embodied Reasoning (ERQA) benchmark, reported to outperform other frontier models on that benchmark.
ARC-AGI-3 (Provider Adapter Harness) scoreIn reporting by testers, Argon scored 97.5% on ARC-AGI-3 when run with a "Provider Adapter Harness" (a test setup that uses vendor-specific API features).
ARC-AGI Standard Harness performanceUnder the ARC-AGI "Standard Harness" (which allows only limited notes between rounds), testers reported Argon dropping to 49.8% at low reasoning; testers also reported API issues that prevented completing all games in that harness.
ARC‑AGI‑3 benchmark result (provider harness)Reported 97.5% on ARC‑AGI‑3 when tested with a vendor‑specific setup ("Provider Adapter Harness"); the report notes the result depends heavily on the testing method and environment.
Benchmark claimsGoogle states that Argon's depth and precision across multi‑step tasks leads it to outperform other frontier models on benchmarks across domains.
Session‑preparation overhead reductionThe accepted optimizations reduced in‑memory session‑preparation overhead by about 40% in long sessions.
Cold package import time reductionThe accepted optimizations reduced cold package import time by 21%.
Artificial Analysis Intelligence Index (v5.0) scoreArgon scored 50 on the Artificial Analysis Intelligence Index v5.0 (reported as Argon's score, with Carbon higher at 59).
Real-to-sim fine-tuning resultGoogle reports synthesizing domain-randomized MuJoCo episodes from one RGB photo to fine-tune a vision-language-action policy, raising success on the corresponding real setup from 56% to 75%.
Benchmark performance (ERQA)Argon scores 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning without helper tools.
Artificial Analysis Intelligence Index score (default thinking level)On Google’s cited Artificial Analysis Intelligence Index at the default “high” thinking level, Argon is reported to score 50 (compared with Hydrogen’s 51; Hydrogen reaches 54 at its highest “xhigh” setting, per Google).
AVAILABILITY
AvailabilityAvailable to Google AI Ultra subscribers and via the paid Gemini API (model id 'gemini-4-argon' in the Gemini API docs).

RELEASE TIMELINE

Published, source-backed release events only.

2 SOURCES · IMPORTANCE 8.0

Google introduces Gemini Robotics 2.5, a more capable vision-language-action model for real-world dexterity

Google announced Gemini Robotics 2.5, an upgraded vision-language-action (VLA) model that increases task breadth, dexterity, reliability, and safety for robots. Google says the model was evaluated on over 200 tasks (showing 30–100% higher success rates and 20–30% faster completion versus prior baselines) and achieved a 96% success rate in a Mercedes‑Benz proof‑of‑concept using Apptronik’s Apollo 2 humanoid on production lines.

→
14 SOURCES · IMPORTANCE 9.0

Google releases Gemini 4 Argon for robotics control, simulation, and policy improvement

Google released Gemini 4 Argon, a general-purpose physical-reasoning model aimed at robot control, Real2Sim scene construction, multimodal labeling of robot video, and iterative policy improvement. Google’s demos show Argon guiding a Boston Dynamics Spot and robot arms in MuJoCo twins, reconstructing scenes from teleoperation video, and synthesizing domain-randomized MuJoCo episodes (reported to raise success from 56% to 75% on one setup); Google reports Argon scores 87% on the ERQA benchmark, and makes it available to Google AI Ultra subscribers and via the paid Gemini API as gemini-4-argon, while publishing only limited per-demo failure data and relying on heavy prompt scaffolding in examples.

→
10 SOURCES · IMPORTANCE 9.0

Google launches Gemini 4 Argon, a new frontier model, globally

Google today made Gemini 4 Argon globally available to developers and enterprises via the API in AI Studio and the Gemini platform, and to Google AI Ultra/Pro subscribers in Antigravity, the Gemini app, and AI Mode in Search. Google positions Argon as a high‑precision frontier model for complex multi‑step tasks across coding, finance, legal, cybersecurity, and science, citing top ranking on the Vals Index, example deployments inside Google (Chrome refactor, ADK optimizations), technical speedups (xTB and Gemma inference via vLLM), and introductory pricing ($2 per million input tokens, $10 per million output tokens, with cached inputs 95% off).

→
2 SOURCES · IMPORTANCE 8.0

Google releases Nano Banana 2.1 image model and halves generation prices

Google released Nano Banana 2.1, an image-generation and editing model running on Gemini 3.6 Flash that the company says improves visual quality, text rendering, character consistency, wide panoramas, and infographics while supporting up to 14 reference images and keeping up to four characters and ten objects consistent. Google also cut Nano Banana pricing roughly in half (for example, 1K images from 6.70¢ to 3.36¢), is rolling 2.1 out across products including the Gemini app, AI Mode in Google Search, Google AI Studio, Flow and Stitch, and will shut down the prior gemini-3.1-flash-image model on October 29, 2026.

→
5 SOURCES · IMPORTANCE 9.0

Google DeepMind unveils Gemini 4 Argon, a frontier model with 1M-token context

Google DeepMind announced Gemini 4 Argon, a new frontier multimodal model optimized for long-horizon reasoning and complex workflows. Argon is rolling out initially to trusted cyber defenders via the Fairwind Program, expands context length to 1 million tokens, reports leading benchmark performance across coding, finance, legal, and video understanding, and will be made more widely available after phased safety testing and engagement with U.S. government pre-release processes; Google also published introductory pricing.

→

VERIFIABLE FACTS

Every value stays attached to a source and date.

CAPABILITIES · Targeted capabilities / domainsDEVELOPER CLAIM

Google describes Argon as built to sustain deep reasoning across complex, long‑horizon workflows and delivering frontier performance in real‑world software engineering, enterprise knowledge work (legal, finance), and cybersecurity defense.

CAPABILITIES · Core capabilities / target workloadsDEVELOPER CLAIM

Argon is engineered for advanced/deep reasoning across complex, long‑horizon workflows and is positioned to excel at heavy‑duty workloads such as coding, cybersecurity defense, enterprise knowledge work (legal, finance), and autonomous cybersecurity patching.

BENCHMARKS · Benchmark reportingDEVELOPER CLAIM

Media reporting cites Google claims that Argon scored higher than competing frontier models (including OpenAI’s GPT‑6 Astra and Anthropic’s Fable/Opus) on a variety of AI benchmarks and is listed as leading on at least one model index cited by Google.

CAPABILITIES · Frontier model capabilities across domainsDEVELOPER CLAIM

Google describes Argon as designed for deep work in complex, real‑world tasks across coding, finance, legal, cybersecurity, and science, with improved multi‑step reasoning, reduced hedging and repetition, and higher precision on long tasks.

BENCHMARKS · Vals Index rankingDEVELOPER CLAIM

Google reports Argon ranks #1 on the Vals Index, an internal measure mentioned as assessing model performance across economic sectors such as finance, coding, legal, and tax work.

BENCHMARKS · Cybersecurity evaluationDEVELOPER CLAIM

Google reports Argon solved all 12 end‑to‑end simulated cyberattack scenarios in their cybersecurity tests (up from five in the previous generation), and describes frontier‑level cyberdefense capabilities intended to help defenders.

LIMITATIONS · Dependence on human oversight with increasing complexityDEVELOPER CLAIM

Across 1,389 internal Argon agent runs, the success rate without human input dropped sharply as task complexity increased, indicating reduced ability to fully automate more complex tasks.

CAPABILITIES · Relative position vs. Gemini capability trendDEVELOPER CLAIM

DeepMind reports Gemini 4 Argon sits slightly above the Gemini capability trend line; the team attributes this position to increased compute and organizational investment rather than AI‑driven acceleration of research.

CAPABILITIES · Claimed capabilities across domainsDEVELOPER CLAIM

Google describes Argon as excelling at deep, multi-step reasoning and structural precision across complex, real-world tasks in coding, finance, legal, cybersecurity, and science; examples include large codebase refactoring, code optimization, finding and patching cybersecurity vulnerabilities, and supporting scientific research.

BENCHMARKS · Embodied Reasoning (ERQA) scoreDEVELOPER CLAIM

Argon scored 87% on the Embodied Reasoning (ERQA) benchmark, reported to outperform other frontier models on that benchmark.

CAPABILITIES · Domains and core strengthsDEVELOPER CLAIM

Described as excelling at deep, multi-step tasks across domains such as coding, finance, legal, cybersecurity, and science; emphasized strengths include physical understanding, code generation, and visual‑semantic reasoning for robotics and real‑world workflows.

CAPABILITIES · Security-research and exploit-development capabilityDEVELOPER CLAIM

Reportedly used by Wiz to find a critical PHP vulnerability; described as especially useful for finding the vulnerability, developing exploits, building attack detection, and validating the patch.

CAPABILITIES · Capabilities (reasoning, coding, tool coordination, long‑horizon workflows)DEVELOPER CLAIM

Google states Gemini 4 Argon can reason, write code, and coordinate tools to execute complex, long‑horizon workflows and deep comprehension tasks.

CAPABILITIES · Capability (large‑data parsing and synthesis)DEVELOPER CLAIM

In an example, Gemini 4 Argon (in Antigravity) parsed more than 80,000 raw files from NASA — including weather logs and engineering, health, and science data across 72 flights on Mars — and organized insights into an executive‑ready deck.

CAPABILITIES · Capability (3D generation and mathematical reasoning)DEVELOPER CLAIM

Google shows Argon producing 5,000 lines of math to render a 3D sunlit library from scratch and then creating a video about the process, demonstrating combined mathematical and 3D‑generation abilities.

CAPABILITIES · Reported domain strengths and use casesDEVELOPER CLAIM

Reported to perform well in areas such as finance, legal, cybersecurity, and science; cited use cases include untangling legacy codebases, accelerating scientific research, robotics, and other domain-specific tasks.

CAPABILITIES · CapabilitiesDEVELOPER CLAIM

Presented as a general-purpose physical-reasoning model for robot control, Real2Sim scene construction (video-to-sim), multimodal labeling of robot video, vision-language-action policies, and iterative policy improvement. Examples in the report include directing a Boston Dynamics Spot, steering an arm with visual coordinates and HTTP motion primitives, and closing a loop with the Gemini Omni video world model.

CAPABILITIES · Reported capabilities (robotics)DEVELOPER CLAIM

Framed as a general-purpose physical-reasoning model for robot control, Real2Sim scene construction, multimodal labeling of robot video, and iterative policy improvement; reported uses include steering simulation, synthesizing domain-randomized episodes, and improving real-world policy success rates.

CAPABILITIES · Core capabilities and typical use casesDEVELOPER CLAIM

Argon is described as a frontier model that "excels at deep, long‑horizon reasoning" and as the ideal choice for the majority of software‑engineering and knowledge‑work tasks.

BENCHMARKS · Artificial Analysis Intelligence Index score (default thinking level)DEVELOPER CLAIM

On Google’s cited Artificial Analysis Intelligence Index at the default “high” thinking level, Argon is reported to score 50 (compared with Hydrogen’s 51; Hydrogen reaches 54 at its highest “xhigh” setting, per Google).

WHAT CHANGED

Stored passport versions, without reconstructed history.

Passport updated58 facts
BENCHMARKS · Artificial Analysis Intelligence Index score (default thinking level)+ On Google’s cited Artificial Analysis Intelligence Index at the default “high” thinking level, Argon is reported to score 50 (compared with Hydrogen’s 51; Hydrogen reaches 54 at its highest “xhigh” setting, per Google).
RELEASE · Launch / availability+ Google describes Gemini 4 Argon as having been launched recently (it preceded the announcement of Gemini 4 Carbon).
CAPABILITIES · Core capabilities and typical use cases+ Argon is described as a frontier model that "excels at deep, long‑horizon reasoning" and as the ideal choice for the majority of software‑engineering and knowledge‑work tasks.
RELEASE · Release− Google released Gemini 4 Argon.
LIMITATIONS · Available thinking‑level settings+ Reporting indicates Argon offers only medium and high thinking levels (contrasting with Hydrogen’s low, medium, high, and new “xhigh” options).
Passport updated55 facts
AVAILABILITY · Availability / Access− Reported available to Google AI Ultra subscribers and through the paid Gemini API; listed with model id 'gemini-4-argon' in the Gemini API models documentation (per the article).
AVAILABILITY · Availability+ Available to Google AI Ultra subscribers and via the paid Gemini API (model id 'gemini-4-argon' in the Gemini API docs).
BENCHMARKS · Benchmark performance (ERQA)+ Argon scores 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning without helper tools.
PRICING · Introductory price (reported)Introductory price reported as $2 per million input tokens and $10 per million output tokens; cached input tokens priced at 95% off input token price.→Reported introductory price: $2 per million input tokens and $10 per million output tokens (stated as the same introductory price as Gemini 4 Argon).
MODALITIES · Modalities− Multimodal: handles robot video, RGB images, vision-language-action control; integrates simulation (MuJoCo) and a video world model (Gemini Omni) in examples.
MODALITIES · Modality+ Described as a multimodal physical-reasoning model combining visual inputs, language, and action/simulation for robotics (video + language + control).
LIMITATIONS · Published metrics limitation+ The Argon robotics post did not publish per-demo failure rates beyond a single VLA case (Google did not publish per-demo failure rates beyond that VLA example).
RELEASE · Release+ Google released Gemini 4 Argon.
RELEASE · Release− Reported released by Google (article published 2026-11-09).
CAPABILITIES · Reported capabilities (robotics)+ Framed as a general-purpose physical-reasoning model for robot control, Real2Sim scene construction, multimodal labeling of robot video, and iterative policy improvement; reported uses include steering simulation, synthesizing domain-randomized episodes, and improving real-world policy success rates.
Passport updated52 facts
AVAILABILITY · Availability / Access+ Reported available to Google AI Ultra subscribers and through the paid Gemini API; listed with model id 'gemini-4-argon' in the Gemini API models documentation (per the article).
AVAILABILITY · Availability− Available to Google AI Ultra subscribers and through the paid Gemini API (listed as model id gemini-4-argon in the Gemini API models docs).
CAPABILITIES · Capabilities+ Presented as a general-purpose physical-reasoning model for robot control, Real2Sim scene construction (video-to-sim), multimodal labeling of robot video, vision-language-action policies, and iterative policy improvement. Examples in the report include directing a Boston Dynamics Spot, steering an arm with visual coordinates and HTTP motion primitives, and closing a loop with the Gemini Omni video world model.
BENCHMARKS · Benchmark (ERQA)Argon scores 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning, which Google reports leads other frontier models on that test.→Reported score of 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning without helper tools; reported by Google as leading other frontier models on that test.
LIMITATIONS · Limitations / evaluation notes+ Reported demos and evaluations rely on long, task-specific prompts, coordinate grids, motion primitives and other scaffolding as part of the system; the article notes Google does not publish per-demo failure rates beyond a single VLA case.
RELEASE · Release+ Reported released by Google (article published 2026-11-09).
RELEASE · Release− Google has released Gemini 4 Argon.
Passport updated50 facts
AVAILABILITY · Availability and distribution channels− Made available to developers, enterprises, scientists, and consumers across Google AI Studio, the Gemini API, Google Cloud, Search, and the Gemini app.
AVAILABILITY · Availability+ Available to Google AI Ultra subscribers and through the paid Gemini API (listed as model id gemini-4-argon in the Gemini API models docs).
SAFETY · Safety testing and rollout approach− Following early safety testing, Gemini 4 Argon was rolled out globally.
BENCHMARKS · ERQA benchmark scoreReported score of 87% on the Embodied Reasoning (ERQA) benchmark, described as testing physical and spatial problem solving in 3D environments.→Argon scores 87% on the Embodied Reasoning Question Answering (ERQA) benchmark using pure reasoning, which Google reports leads other frontier models on that test.
CAPABILITIES · Capabilities+ Framed as a general-purpose physical-reasoning model for robot control, Real2Sim scene construction, multimodal labeling of robot video, and iterative policy improvement rather than as a single end-effector policy.
RELEASE · Release and rollout− Rolled out globally in October 2026 as a frontier reasoning model.
MODALITIES · Modalities+ Multimodal: handles robot video, RGB images, vision-language-action control; integrates simulation (MuJoCo) and a video world model (Gemini Omni) in examples.
BENCHMARKS · Real-to-sim fine-tuning result+ Google reports synthesizing domain-randomized MuJoCo episodes from one RGB photo to fine-tune a vision-language-action policy, raising success on the corresponding real setup from 56% to 75%.
RELEASE · Release+ Google has released Gemini 4 Argon.
MODALITIES · Modality — physical‑world reasoning / robotics− Powers physical‑world reasoning through a robotics variant called 'Gemini 4 Argon for Robotics'.
LIMITATIONS · Limitations / evaluation reporting+ Google's robotics examples use long, task-specific prompts, coordinate grids, motion primitives, and verifiers (scaffolding is part of the system under test); Google does not publish per-demo failure rates beyond the single VLA case.
SAFETY · Practitioner concerns about simulation fidelity+ Early practitioner reaction highlighted concerns that simulation twins can look visually convincing while object/contact dynamics behave incorrectly, which matters for manipulation performance.
Passport updated47 facts
AVAILABILITY · Availability and distribution channels+ Made available to developers, enterprises, scientists, and consumers across Google AI Studio, the Gemini API, Google Cloud, Search, and the Gemini app.
AVAILABILITY · Availability and early usage− Available on Google's developer platforms; in the first week after launch, usage ramped up 3.5× faster than Gemini 3 (reported by Google during earnings remarks).
SAFETY · Dual-use / offensive potential− Reports indicate the model's capabilities were used to develop exploits, highlighting its dual-use potential where the same abilities that help find vulnerabilities can also assist attackers.
SAFETY · Safety testing and rollout approach+ Following early safety testing, Gemini 4 Argon was rolled out globally.
RELEASE · Release and rollout+ Rolled out globally in October 2026 as a frontier reasoning model.
RELEASE · Launch timing− Launched two weeks before Gemini 4 Carbon (Carbon announced Oct 28, 2026).
MODALITIES · Modality− Described by Google as multimodal; the blog highlights Argon’s visual understanding alongside reasoning and other modalities.
MODALITIES · Modality — physical‑world reasoning / robotics+ Powers physical‑world reasoning through a robotics variant called 'Gemini 4 Argon for Robotics'.
CAPABILITIES · Capabilities / target use cases+ Positioned for complex coding, financial analysis, and legal workflows.
Passport updated46 facts
BENCHMARKS · Artificial Analysis Intelligence Index (v5.0) score+ Argon scored 50 on the Artificial Analysis Intelligence Index v5.0 (reported as Argon's score, with Carbon higher at 59).
AVAILABILITY · Availability and early usage+ Available on Google's developer platforms; in the first week after launch, usage ramped up 3.5× faster than Gemini 3 (reported by Google during earnings remarks).
CAPABILITIES · Reported domain strengths and use cases+ Reported to perform well in areas such as finance, legal, cybersecurity, and science; cited use cases include untangling legacy codebases, accelerating scientific research, robotics, and other domain-specific tasks.
AVAILABILITY · Availability / integration− Google demonstrates Gemini 4 Argon in the Gemini app and in Antigravity/AI Studio prototypes and workflows, indicating integration into those Google demonstration environments.
PRICING · Introductory pricing+ Introductory price reported as $2 per million input tokens and $10 per million output tokens; cached input tokens priced at 95% off input token price.
PRICING · Introductory pricing− Introduced at an introductory price of $2 per million input tokens and $10 per million output tokens; Google notes a roughly 95% discount on cached input tokens; press reports say Google may raise prices after the introductory period.
RELEASE · Launch and availability (global)− Google and DeepMind announced Gemini 4 Argon as globally available at launch for developers, enterprises, and Google AI Pro/Ultra subscribers (October 2026 announcements).
RELEASE · Launch timing+ Launched two weeks before Gemini 4 Carbon (Carbon announced Oct 28, 2026).
Passport updated44 facts
CAPABILITIES · Capability (3D generation and mathematical reasoning)+ Google shows Argon producing 5,000 lines of math to render a 3D sunlit library from scratch and then creating a video about the process, demonstrating combined mathematical and 3D‑generation abilities.
AVAILABILITY · Availability / integration+ Google demonstrates Gemini 4 Argon in the Gemini app and in Antigravity/AI Studio prototypes and workflows, indicating integration into those Google demonstration environments.
CAPABILITIES · Capability (large‑data parsing and synthesis)+ In an example, Gemini 4 Argon (in Antigravity) parsed more than 80,000 raw files from NASA — including weather logs and engineering, health, and science data across 72 flights on Mars — and organized insights into an executive‑ready deck.
MODALITIES · Modality+ Described by Google as multimodal; the blog highlights Argon’s visual understanding alongside reasoning and other modalities.
MODALITIES · Modalities− Examples in the post show Argon working with visual inputs and code: generating 3D simulations from single images, integrating with Blender and Three.js, turning dense data into interactive visuals, and producing interactive 3D experiences—indicating multimodal capabilities (vision + code/3D output).
AVAILABILITY · Access paths and product availability− Available via the API in AI Studio and the Gemini platform; available to Google AI Ultra subscribers in Google Antigravity; available to Google AI Ultra and Pro subscribers in the Gemini app and AI Mode in Search; developers can build using the paid tier of the Gemini API and related developer tools.
CAPABILITIES · Capabilities (reasoning, coding, tool coordination, long‑horizon workflows)+ Google states Gemini 4 Argon can reason, write code, and coordinate tools to execute complex, long‑horizon workflows and deep comprehension tasks.
CAPABILITIES · Capability (task decomposition and planning for simulation/robotics)+ A research experiment reported that a Gemini 4 Argon model broke down tasks into steps that the Genie 4 world model then simulated in real time, illustrating Argon’s use for task decomposition and planning in simulation/robotics contexts.
Passport updated40 facts
CAPABILITIES · Handled large Python codebase+ Argon operated on a codebase spanning over 200,000 lines of Python across approximately 800 files.
BENCHMARKS · Cold package import time reduction+ The accepted optimizations reduced cold package import time by 21%.
CAPABILITIES · Optimizations produced during unattended run+ An 11‑hour unattended run of Gemini 4 Argon produced two accepted optimizations for Google Agent Development Kit's codebase.
BENCHMARKS · Session‑preparation overhead reduction+ The accepted optimizations reduced in‑memory session‑preparation overhead by about 40% in long sessions.
CAPABILITIES · Codebase processing and testing+ Argon read repository guidelines, established performance baselines, proposed optimizations, reviewed its own changes, and ran all 16,892 open‑source unit tests for the repository.