--- title: Gemini 4 Argon (In-depth benchmark analysis): Did Google... description: Google announced Gemini 4 Argon, and independent benchmarks show mixed results: it leads Zapier's automation bench, Serge's spreadsheet benchmark, and the... canonical: https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback? | daily.dev og:description: Google announced Gemini 4 Argon, and independent benchmarks show mixed results: it leads Zapier's automation bench, Serge's spreadsheet benchmark, and the... og:url: https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j og:image: https://api.daily.dev/og/posts/BvLWvbY7J.png og:image:alt: Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback? og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback? **[AICodeKing](https://daily.dev/sources/aicodeking)** · 14 min read · 0 upvotes · 0 comments ## Summary Google announced Gemini 4 Argon, and independent benchmarks show mixed results: it leads Zapier's automation bench, Serge's spreadsheet benchmark, and the Val's index, and shows the lowest hallucination rate among top models (though with lower knowledge accuracy than Astra). Coding results are inconsistent — strong on DeepSWE but behind Opus and Astra on Frontier Sway and terminal bench. Pricing has an introductory discount ($2/$10 per million tokens) that doubles to $4/$20 after an unspecified period, changing the cost-value comparison significantly versus Claude Opus, Sonnet, and GPT 6.1. Initial access is limited to Google's Fairwind program and select partners, with no public release date given. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How does Gemini 4 Argon's pricing change after the introductory period? Google advertises an introductory rate of $2 per million input tokens and $10 per million output tokens with a 95% discount on cached input, but a footnote states these prices double to $4 and $20 after the introductory period ends, with no exact end date given. At the promotional rate, artificial analysis measures about $1.99 per intelligence index task, rising to roughly $3.98 after the discount expires. _Budgeting an LLM-powered app means tracking pricing changes like Argon's before they hit your bill on daily.dev._ ### How does Gemini 4 Argon compare to Claude Opus 5.5 and Sonnet 5.5 on coding benchmarks? Results are mixed: Argon leads on DeepSWE v1.1 with 77.9% versus Opus's 74.2%, but trails on Frontier Sway v2 (55% versus Opus's 62.3% and Astra's 65.5%) and Google's own terminal bench 4 (57.4% versus Opus's 66.4%). On Val's app building and code migration benchmarks, Argon ranks second behind Sonnet 5.5, though the app-building gap is under half a percentage point, within standard error. _Comparing coding agents before switching workflows is exactly the kind of decision daily.dev readers track closely._ ### What is Gemini 4 Argon's hallucination rate compared to other models? Artificial analysis reports a 15% hallucination rate on AA omniscience for Argon, the lowest among models scoring at least 45 on its intelligence index, compared to 51% for Astra at max. However, Argon's knowledge accuracy is only 50% versus Astra's 63%, meaning it answers fewer questions correctly but more often admits uncertainty rather than giving a confident wrong answer. _Understanding a model's tendency to admit uncertainty matters for anyone choosing an AI assistant for research-heavy work._ --- Tags: [#ai](https://daily.dev/tags/ai), [#google](https://daily.dev/tags/google), [#ai-agents](https://daily.dev/tags/ai-agents), [#claude](https://daily.dev/tags/claude), [#google-gemini](https://daily.dev/tags/google-gemini) [View this post on daily.dev](https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback?","url":"https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j"},"datePublished":"2026-10-01T09:28:33.476Z","dateModified":"2026-10-01T09:29:24.679Z","description":"Google announced Gemini 4 Argon, and independent benchmarks show mixed results: it leads Zapier's automation bench, Serge's spreadsheet benchmark, and the...","image":"https://i.ytimg.com/vi/ctP3nhMwsEM/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/ctP3nhMwsEM/sddefault.jpg","isAccessibleForFree":true,"articleSection":"AICodeKing","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"AICodeKing","logo":"https://media.daily.dev/image/upload/s--x7nDUfWj--/f_auto,q_auto/v1768208312/logos/aicodeking","url":"https://daily.dev/sources/aicodeking"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,google,ai-agents,claude,google-gemini","timeRequired":"PT14M","video":{"@type":"VideoObject","name":"Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback?","description":"Google announced Gemini 4 Argon, and independent benchmarks show mixed results: it leads Zapier's automation bench, Serge's spreadsheet benchmark, and the...","thumbnailUrl":"https://i.ytimg.com/vi/ctP3nhMwsEM/sddefault.jpg","uploadDate":"2026-10-01T09:28:33.476Z","duration":"PT14M","url":"https://api.daily.dev/r/BvLWvbY7J","embedUrl":"https://www.youtube.com/embed/ctP3nhMwsEM"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"AICodeKing","item":"https://daily.dev/sources/aicodeking"},{"@type":"ListItem","position":3,"name":"Gemini 4 Argon (In-depth benchmark analysis): Did Google make a big comeback?"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/gemini-4-argon-in-depth-benchmark-analysis-did-google-make-a-big-comeback--bvlwvby7j#faq","mainEntity":[{"@type":"Question","name":"How does Gemini 4 Argon's pricing change after the introductory period?","acceptedAnswer":{"@type":"Answer","text":"Google advertises an introductory rate of $2 per million input tokens and $10 per million output tokens with a 95% discount on cached input, but a footnote states these prices double to $4 and $20 after the introductory period ends, with no exact end date given. At the promotional rate, artificial analysis measures about $1.99 per intelligence index task, rising to roughly $3.98 after the discount expires. Budgeting an LLM-powered app means tracking pricing changes like Argon's before they hit your bill on daily.dev."}},{"@type":"Question","name":"How does Gemini 4 Argon compare to Claude Opus 5.5 and Sonnet 5.5 on coding benchmarks?","acceptedAnswer":{"@type":"Answer","text":"Results are mixed: Argon leads on DeepSWE v1.1 with 77.9% versus Opus's 74.2%, but trails on Frontier Sway v2 (55% versus Opus's 62.3% and Astra's 65.5%) and Google's own terminal bench 4 (57.4% versus Opus's 66.4%). On Val's app building and code migration benchmarks, Argon ranks second behind Sonnet 5.5, though the app-building gap is under half a percentage point, within standard error. Comparing coding agents before switching workflows is exactly the kind of decision daily.dev readers track closely."}},{"@type":"Question","name":"What is Gemini 4 Argon's hallucination rate compared to other models?","acceptedAnswer":{"@type":"Answer","text":"Artificial analysis reports a 15% hallucination rate on AA omniscience for Argon, the lowest among models scoring at least 45 on its intelligence index, compared to 51% for Astra at max. However, Argon's knowledge accuracy is only 50% versus Astra's 63%, meaning it answers fewer questions correctly but more often admits uncertainty rather than giving a confident wrong answer. Understanding a model's tendency to admit uncertainty matters for anyone choosing an AI assistant for research-heavy work."}}]} ```