--- title: The AI model that just scored 65% on DeepSWE isn’t the... description: Google launched Gemini 3.7 Flash as its new coding-agent and automation model, arriving just three weeks after Gemini 3.6 Flash. Benchmark scores jumped... canonical: https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The AI model that just scored 65% on DeepSWE isn’t the one Google promised. | daily.dev og:description: Google launched Gemini 3.7 Flash as its new coding-agent and automation model, arriving just three weeks after Gemini 3.6 Flash. Benchmark scores jumped... og:url: https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk og:image: https://api.daily.dev/og/posts/bmcMtv1nK.png og:image:alt: The AI model that just scored 65% on DeepSWE isn’t the one Google promised. 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. # The AI model that just scored 65% on DeepSWE isn’t the one Google promised. **[The New Stack](https://daily.dev/sources/newstack)** · 5 min read · 0 upvotes · 0 comments ## Summary Google launched Gemini 3.7 Flash as its new coding-agent and automation model, arriving just three weeks after Gemini 3.6 Flash. Benchmark scores jumped notably: DeepSWE v1.1 rose from 49% to 65.3% and FrontierCode 1.1 Main from 34.4% to 43.6%. The model introduces a thinking_level setting (low/medium/high) replacing thinking_budget, and API pricing is temporarily halved to $0.75/$3.75 per million input/output tokens, doubling to $1.50/$7.50 on January 1, 2027. Migrating from Gemini 3.5 Flash, 3 Flash Preview, or 3.1 Pro requires removing temperature, top_p, top_k, candidate_count, and prefilled model turns, and adopting previous_interaction_id. Automation benchmarks also improved (GDP.pdf 22%→34%, AutomationBench 17%→30.4%). Gemini 3.5 Pro, promised for June, remains unreleased while Google has begun training Gemini 4. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What changed in Gemini 3.7 Flash compared to Gemini 3.6 Flash for coding agents? Gemini 3.7 Flash scored 65.3% on DeepSWE v1.1, up from 49% for Gemini 3.6 Flash, and 43.6% on FrontierCode 1.1 Main, up from 34.4%. Its WebDev Arena score rose from 1,538 to 1,588. Google says it's less likely to get stuck when something goes wrong and better at knowing when it needs more information before proceeding. _Teams weighing an agent model upgrade can track benchmark shifts like these on daily.dev._ ### When does the Gemini 3.7 Flash API price increase and by how much? Gemini 3.7 Flash currently costs $0.75 per million input tokens and $3.75 per million output tokens, a rate also applied to Gemini 3.6 Flash until December 31. On January 1, 2027, both rates double to $1.50 per million input tokens and $7.50 per million output tokens, so teams building on the introductory price need to budget for the increase. _Teams budgeting agent workloads around API pricing can watch for changes like this on daily.dev._ ### What do I need to change to migrate to Gemini 3.7 Flash's thinking_level API? Migrating from Gemini 3.5 Flash, Gemini 3 Flash Preview, or Gemini 3.1 Pro requires removing temperature, top_p, and top_k, replacing the numeric thinking_budget with the new thinking_level setting, and dropping candidate_count. Developers must also remove prefilled model turns and standardize multi-turn interactions around the server-side previous_interaction_id, with every FunctionResponse including its call ID and function name. _Developers planning an AI model migration can track breaking API changes like this on daily.dev._ ## Similar posts on daily.dev - [Google's Gemini 3.5 Flash costs 3x the model it replaced, and the era of cheap AI is ending](https://daily.dev/posts/google-s-gemini-3-5-flash-costs-3x-the-model-it-replaced-and-the-era-of-cheap-ai-is-ending-ochkeykzb) · XDA Developers · 2 upvotes · 0 comments - [Google releases Gemini 3 Flash, enabling faster, more cost effective reasoning](https://daily.dev/posts/google-releases-gemini-3-flash-enabling-faster-more-cost-effective-reasoning-1y9yu2yfb) · SD Times · 0 upvotes · 0 comments - [Gemini 3.1 Flash-Lite: Built for intelligence at scale](https://daily.dev/posts/gemini-3-1-flash-lite-built-for-intelligence-at-scale-oz1mpgrn8) · DeepMind · 1 upvotes · 1 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#google](https://daily.dev/tags/google), [#ai-agents](https://daily.dev/tags/ai-agents), [#google-gemini](https://daily.dev/tags/google-gemini) [View this post on daily.dev](https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk) ```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":"The AI model that just scored 65% on DeepSWE isn’t the one Google promised.","url":"https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk"},"datePublished":"2026-08-13T20:02:00.363Z","dateModified":"2026-08-14T00:33:36.437Z","description":"Google launched Gemini 3.7 Flash as its new coding-agent and automation model, arriving just three weeks after Gemini 3.6 Flash. Benchmark scores jumped...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/2781be17f415252b9c56922fe1826289?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/2781be17f415252b9c56922fe1826289?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The New Stack","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":"The New Stack","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/newstack","url":"https://daily.dev/sources/newstack"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,google,ai-agents,google-gemini","timeRequired":"PT5M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The New Stack","item":"https://daily.dev/sources/newstack"},{"@type":"ListItem","position":3,"name":"The AI model that just scored 65% on DeepSWE isn’t the one Google promised."}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-ai-model-that-just-scored-65-on-deepswe-isn-t-the-one-google-promised--bmcmtv1nk#faq","mainEntity":[{"@type":"Question","name":"What changed in Gemini 3.7 Flash compared to Gemini 3.6 Flash for coding agents?","acceptedAnswer":{"@type":"Answer","text":"Gemini 3.7 Flash scored 65.3% on DeepSWE v1.1, up from 49% for Gemini 3.6 Flash, and 43.6% on FrontierCode 1.1 Main, up from 34.4%. Its WebDev Arena score rose from 1,538 to 1,588. Google says it's less likely to get stuck when something goes wrong and better at knowing when it needs more information before proceeding. Teams weighing an agent model upgrade can track benchmark shifts like these on daily.dev."}},{"@type":"Question","name":"When does the Gemini 3.7 Flash API price increase and by how much?","acceptedAnswer":{"@type":"Answer","text":"Gemini 3.7 Flash currently costs $0.75 per million input tokens and $3.75 per million output tokens, a rate also applied to Gemini 3.6 Flash until December 31. On January 1, 2027, both rates double to $1.50 per million input tokens and $7.50 per million output tokens, so teams building on the introductory price need to budget for the increase. Teams budgeting agent workloads around API pricing can watch for changes like this on daily.dev."}},{"@type":"Question","name":"What do I need to change to migrate to Gemini 3.7 Flash's thinking_level API?","acceptedAnswer":{"@type":"Answer","text":"Migrating from Gemini 3.5 Flash, Gemini 3 Flash Preview, or Gemini 3.1 Pro requires removing temperature, top_p, and top_k, replacing the numeric thinking_budget with the new thinking_level setting, and dropping candidate_count. Developers must also remove prefilled model turns and standardize multi-turn interactions around the server-side previous_interaction_id, with every FunctionResponse including its call ID and function name. Developers planning an AI model migration can track breaking API changes like this on daily.dev."}}]} ```