[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"article-gemini-3-7-flash-launch-coding-gains-en":3,"article-related-gemini-3-7-flash-launch-coding-gains-en":29,"series-model-release-342db385-c34c-4a62-8263-4a2237105dbe":74},{"id":4,"slug":5,"title":6,"content":7,"summary":8,"source":9,"source_url":10,"author":11,"image_url":12,"cover_image":12,"category":13,"language":14,"translated_content":11,"related_article_id":15,"keywords":16,"key_takeaways":22,"views":26,"created_at":27,"published_at":28,"topic_cluster_id":11},"342db385-c34c-4a62-8263-4a2237105dbe","gemini-3-7-flash-launch-coding-gains-en","Gemini 3.7 Flash arrives with faster coding gains","\u003Cp data-speakable=\"summary\">\u003Ca href=\"\u002Ftag\u002Fgoogle\">Google\u003C\u002Fa> released \u003Ca href=\"\u002Ftag\u002Fgemini\">Gemini\u003C\u002Fa> 3.7 Flash three weeks after its last model, with stronger coding and workflow scores.\u003C\u002Fp>\u003Cp>Google is shipping \u003Ca href=\"https:\u002F\u002Fai.google.dev\u002Fgemini-api\u002Fdocs\u002Fmodels\" target=\"_blank\" rel=\"noopener\">Gemini 3.7 Flash\u003C\u002Fa> just three weeks after the previous release, and the timing says a lot about how fast the company wants this line to move. The new model lands with better \u003Ca href=\"\u002Ftag\u002Fbenchmark\">benchmark\u003C\u002Fa> numbers for coding, web development, and document-heavy work, plus a lower launch price through the end of 2026.\u003C\u002Fp>\u003Cp>That combination matters because Flash is the model tier developers are most likely to use in production when cost, latency, and quality all matter at once. Google is also pushing it into the \u003Ca href=\"https:\u002F\u002Fgemini.google.com\" target=\"_blank\" rel=\"noopener\">Gemini app\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Faistudio.google.com\" target=\"_blank\" rel=\"noopener\">Google AI Studio\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.android.com\u002Fstudio\u002F\" target=\"_blank\" rel=\"noopener\">Android Studio\u003C\u002Fa>, and its enterprise tools, so this is not a lab-only update.\u003C\u002Fp>\u003Ctable>\u003Cthead>\u003Ctr>\u003Cth>Metric\u003C\u002Fth>\u003Cth>Gemini 3.6 Flash\u003C\u002Fth>\u003Cth>Gemini 3.7 Flash\u003C\u002Fth>\u003C\u002Ftr>\u003C\u002Fthead>\u003Ctbody>\u003Ctr>\u003Ctd>DeepSWE v1.1\u003C\u002Ftd>\u003Ctd>49.0%\u003C\u002Ftd>\u003Ctd>65.3%\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>FrontierCode 1.1 Main\u003C\u002Ftd>\u003Ctd>34.4%\u003C\u002Ftd>\u003Ctd>43.6%\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>WebDev Arena Elo\u003C\u002Ftd>\u003Ctd>1538\u003C\u002Ftd>\u003Ctd>1588\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>GDP.pdf\u003C\u002Ftd>\u003Ctd>22.0%\u003C\u002Ftd>\u003Ctd>34.0%\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>AutomationBench\u003C\u002Ftd>\u003Ctd>17.0%\u003C\u002Ftd>\u003Ctd>30.4%\u003C\u002Ftd>\u003C\u002Ftr>\u003Ctr>\u003Ctd>Launch price per 1M tokens\u003C\u002Ftd>\u003Ctd>$1.50 input \u002F $7.50 output\u003C\u002Ftd>\u003Ctd>$0.75 input \u002F $3.75 output\u003C\u002Ftd>\u003C\u002Ftr>\u003C\u002Ftbody>\u003C\u002Ftable>\u003Ch2>Google is moving Flash on a short clock\u003C\u002Fh2>\u003Cp>The biggest story here is cadence. Google says Gemini 3.7 Flash is a direct result of developer feedback and algorithmic changes, and it arrives only three weeks after the last model. That is a much faster rhythm than the old “wait months for a big jump” model that many teams were used to.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1786667569978-4y4d.png\" alt=\"Gemini 3.7 Flash arrives with faster coding gains\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>In practice, that means developers have to pay attention to the Flash line the way they would watch a fast-moving open-source project. Model behavior, pricing, and deployment targets can shift quickly, and teams that build on Google’s stack may need to test updates more often.\u003C\u002Fp>\u003Cp>Google says the model brings substantial improvements in software engineering, web development, and knowledge work. The phrasing is broad, but the numbers are more useful than the marketing copy.\u003C\u002Fp>\u003Cul>\u003Cli>DeepSWE v1.1 rises from 49.0% to 65.3%\u003C\u002Fli>\u003Cli>FrontierCode 1.1 Main climbs from 34.4% to 43.6%\u003C\u002Fli>\u003Cli>WebDev Arena Elo moves from 1538 to 1588\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Those gains suggest Google is focusing on practical developer tasks instead of chasing abstract benchmark bragging rights. Debugging, issue resolution, layout generation, and instruction following are the kinds of jobs that show up in real product work every day.\u003C\u002Fp>\u003Ch2>The coding gains are the clearest signal\u003C\u002Fh2>\u003Cp>Google says Gemini 3.7 Flash shows strong gains over 3.6 Flash for debugging and issue resolution. That matters because debugging is where smaller misses become expensive fast: a model that can identify the right fix with fewer retries saves time, tokens, and human attention.\u003C\u002Fp>\u003Cp>For web development, Google says the model generates more functional layouts and feature-complete apps in fewer prompts. That is a meaningful claim for teams using AI to scaffold front ends, because “fewer prompts” often means fewer half-finished outputs and less cleanup.\u003C\u002Fp>\u003Cblockquote>\u003Cp>“This new model is a direct result of developer feedback and algorithmic innovations that [Google looks] forward to bringing to future models.”\u003C\u002Fp>\u003Cfooter>Google, via the Gemini 3.7 Flash announcement\u003C\u002Ffooter>\u003C\u002Fblockquote>\u003Cp>That line matters because it frames Flash as a feedback loop, not a one-off release. Google is using developer usage data to shape the next version, then shipping the result quickly enough that teams can feel the changes while the product cycle is still fresh.\u003C\u002Fp>\u003Ch2>Business workflows and document work improved too\u003C\u002Fh2>\u003Cp>Google is also pitching Gemini 3.7 Flash for finance, law, biosciences, and other knowledge-dense fields. The most concrete example is the GDP.pdf benchmark, where the model goes from 22.0% to 34.0%.\u003C\u002Fp>\n\u003Cfigure class=\"my-6\">\u003Cimg src=\"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1786667568442-azhs.png\" alt=\"Gemini 3.7 Flash arrives with faster coding gains\" class=\"rounded-xl w-full\" loading=\"lazy\" \u002F>\u003C\u002Ffigure>\n\u003Cp>AutomationBench shows a similar jump, from 17.0% to 30.4%. That is the kind of change that matters for business workflows, where the model has to follow instructions, read documents, and keep going when the task gets messy.\u003C\u002Fp>\u003Cul>\u003Cli>GDP.pdf: 34.0% versus 22.0%\u003C\u002Fli>\u003Cli>AutomationBench: 30.4% versus 17.0%\u003C\u002Fli>\u003Cli>Launch pricing through 2026: $0.75 input and $3.75 output per 1M tokens\u003C\u002Fli>\u003Cli>Gemini app rollout: Spark tier first, with AI Pro and Ultra required\u003C\u002Fli>\u003C\u002Ful>\u003Cp>Google also says the model better adapts to roadblocks, clarifies intent when needed, and follows instructions with greater fidelity. In plain English, that means fewer dead ends when the model is asked to do multi-step work across tools and documents.\u003C\u002Fp>\u003Cp>The company says it is also adding updated safeguards for Chemical, Biological, Radiological, and Nuclear misuse and cyber offense, while keeping beneficial use cases enabled under its bioresilience and cyber programs. That matters because the more capable these models get at tool use and planning, the more important the guardrails become.\u003C\u002Fp>\u003Ch2>Price and placement make this release easy to notice\u003C\u002Fh2>\u003Cp>The pricing is probably the other headline. Until the end of 2026, Gemini 3.7 Flash costs $0.75 per 1 million input tokens and $3.75 per 1 million output tokens, which is half the launch price of the previous model.\u003C\u002Fp>\u003Cp>That makes the model easier to trial in real products, especially for teams that have been waiting for a price point that fits frequent use. It also puts more pressure on rivals like \u003Ca href=\"https:\u002F\u002Fopenai.com\" target=\"_blank\" rel=\"noopener\">OpenAI\u003C\u002Fa>, \u003Ca href=\"https:\u002F\u002Fwww.anthropic.com\" target=\"_blank\" rel=\"noopener\">Anthropic\u003C\u002Fa>, and \u003Ca href=\"https:\u002F\u002Fwww.meta.ai\" target=\"_blank\" rel=\"noopener\">Meta\u003C\u002Fa> to keep their own price-performance story competitive.\u003C\u002Fp>\u003Cp>Google is not hiding the rollout either. Gemini 3.7 Flash is already rolling into Spark in the \u003Ca href=\"https:\u002F\u002Fgemini.google.com\" target=\"_blank\" rel=\"noopener\">Gemini app\u003C\u002Fa>, with \u003Ca href=\"https:\u002F\u002Fcloud.google.com\u002Fvertex-ai\u002Fgenerative-ai\u002Fdocs\" target=\"_blank\" rel=\"noopener\">Gemini Enterprise Agent Platform\u003C\u002Fa> and the Gemini Enterprise app on the list too. It is also available in \u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fgoogle-gemini\u002Fgemini-cli\" target=\"_blank\" rel=\"noopener\">Google Antigravity\u003C\u002Fa> and \u003Ca href=\"https:\u002F\u002Fdeveloper.android.com\u002Fstudio\u002Fpreview\u002Fgemini\" target=\"_blank\" rel=\"noopener\">Android Studio\u003C\u002Fa>.\u003C\u002Fp>\u003Cp>For developers, the practical takeaway is simple: this is a model update worth benchmarking sooner rather than later. If your app uses Flash for coding help, document processing, or agentic workflows, the combination of better scores and lower pricing may change your cost profile enough to justify another round of tests.\u003C\u002Fp>\u003Cp>Google’s pace suggests the Flash family will keep changing quickly, and the next question is whether these gains hold up once more teams push the model into real production traffic. If they do, the companies that test early will have the clearest edge in cost and output quality.\u003C\u002Fp>","Google released Gemini 3.7 Flash three weeks after its last model, with stronger coding, web dev, and business workflow scores.","9to5google.com","https:\u002F\u002F9to5google.com\u002F2026\u002F08\u002F13\u002Fgemini-3-7-flash-launch\u002F",null,"https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1786667569978-4y4d.png","model-release","en","93e19c1d-1ffc-4f12-a336-f39afb16e8ac",[17,18,19,20,21],"Gemini 3.7 Flash","Google AI","coding benchmarks","web development","token pricing",[23,24,25],"Gemini 3.7 Flash launches three weeks after the previous model.","Benchmark gains are strongest in coding, web dev, and document workflows.","Google cut launch pricing to half the previous model and is rolling it into multiple products.",1,"2026-08-14T00:32:29.223591+00:00","2026-08-14T00:32:29.216+00:00",{"tags":30,"relatedLang":33,"relatedPosts":37},[31],{"name":18,"slug":32},"google-ai",{"id":15,"slug":34,"title":35,"language":36},"gemini-3-7-flash-launch-coding-gains-zh","Gemini 3.7 Flash 提速上線，寫程式更強","zh",[38,44,50,56,62,68],{"id":39,"slug":40,"title":41,"cover_image":42,"image_url":42,"created_at":43,"category":13},"ba2f9faf-d709-4570-b835-5e4450aa7fce","august-2026-model-rankings-claude-kimi-en","August 2026 model rankings: Claude leads text, Kimi coding","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1786473195961-dj0x.png","2026-08-11T18:32:47.781195+00:00",{"id":45,"slug":46,"title":47,"cover_image":48,"image_url":48,"created_at":49,"category":13},"3531feda-a023-4b38-9f1c-69f6b1a37f8e","qwen38-max-top-tier-claude-comparison-en","Qwen3.8-Max pushes Alibaba into the top tier","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1786149191019-ls21.png","2026-08-08T00:32:53.612458+00:00",{"id":51,"slug":52,"title":53,"cover_image":54,"image_url":54,"created_at":55,"category":13},"ff8312eb-e6ac-4bbc-85d5-38844a1c1964","qwen38-max-agentic-work-real-frontier-en","Qwen3.8-Max proves that agentic work is the real frontier","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1785844979892-cdc1.png","2026-08-04T12:02:32.157065+00:00",{"id":57,"slug":58,"title":59,"cover_image":60,"image_url":60,"created_at":61,"category":13},"ccdf0c22-70d3-474f-9fa1-592694c564b1","google-earth-should-not-ship-ai-image-generation-en","Google Earth Should Not Ship AI Image Generation","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1785805372500-cnzw.png","2026-08-04T01:02:30.317498+00:00",{"id":63,"slug":64,"title":65,"cover_image":66,"image_url":66,"created_at":67,"category":13},"6e9aa97d-d130-4c68-a2cd-d7bd78b7c610","try-claude-opus-4-7-benchmarks-safety-en","Try Claude Opus 4.7 and read its benchmarks","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1785720761463-1tp9.png","2026-08-03T01:32:20.969744+00:00",{"id":69,"slug":70,"title":71,"cover_image":72,"image_url":72,"created_at":73,"category":13},"40da5e56-c978-4c19-b039-71d4121a46eb","opus-5-cut-cost-without-losing-quality-en","Opus 5 lets you cut cost without losing quality","https:\u002F\u002Fxxdpdyhzhpamafnrdkyq.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Fcovers\u002Finline-1785607382907-76ym.png","2026-08-01T18:02:40.854976+00:00",[75,80,85,90,95,100,105,110,115,120],{"id":76,"slug":77,"title":78,"created_at":79},"d4cffde7-9b50-4cc7-bb68-8bc9e3b15477","nvidia-rubin-ai-supercomputer-en","NVIDIA Unveils Rubin: A Leap in AI Supercomputing","2026-03-25T16:24:35.155565+00:00",{"id":81,"slug":82,"title":83,"created_at":84},"eab919b9-fbac-4048-89fc-afad6749ccef","google-gemini-ai-innovations-2026-en","Google's AI Leap with Gemini Innovations in 2026","2026-03-25T16:27:18.841838+00:00",{"id":86,"slug":87,"title":88,"created_at":89},"5f5cfc67-3384-4816-a8f6-19e44d90113d","gap-google-gemini-ai-checkout-en","Gap Teams Up with Google Gemini for AI-Driven Checkout","2026-03-25T16:27:46.483272+00:00",{"id":91,"slug":92,"title":93,"created_at":94},"f6d04567-47f6-49ec-804c-52e61ab91225","ai-model-release-wave-march-2026-en","Navigating the AI Model Release Wave of March 2026","2026-03-25T16:28:45.409716+00:00",{"id":96,"slug":97,"title":98,"created_at":99},"895c150c-569e-4fdf-939d-dade785c990e","small-language-models-transform-ai-en","Small Language Models: Llama 3.2 and Phi-3 Transform AI","2026-03-25T16:30:26.688313+00:00",{"id":101,"slug":102,"title":103,"created_at":104},"38eb1d26-d961-4fd3-ae12-9c4089680f5f","midjourney-v8-alpha-features-pricing-en","Midjourney V8 Alpha: A Deep Dive into Its Features and Pricing","2026-03-26T01:25:36.387587+00:00",{"id":106,"slug":107,"title":108,"created_at":109},"bf36bb9e-3444-4fb8-ab19-0df6bc9d8271","rag-2026-indispensable-ai-bridge-en","RAG in 2026: The Indispensable AI Bridge","2026-03-26T01:28:34.472046+00:00",{"id":111,"slug":112,"title":113,"created_at":114},"60881d6d-2310-44ef-b1fb-7f98e9dd2f0e","xiaomi-mimo-trio-agents-robots-voice-en","Xiaomi’s MiMo trio targets agents, robots, and voice","2026-03-28T03:05:08.899895+00:00",{"id":116,"slug":117,"title":118,"created_at":119},"f063d8d1-41d1-4de4-8ebc-6c40511b9369","xiaomi-mimo-v2-pro-1t-moe-agents-en","Xiaomi MiMo-V2-Pro: 1T MoE Model for Agents","2026-03-28T03:06:19.238032+00:00",{"id":121,"slug":122,"title":123,"created_at":124},"a1379e9a-6785-4ff5-9b0a-8cff55f8264f","cursor-composer-2-started-from-kimi-en","Cursor’s Composer 2 started from Kimi","2026-03-28T03:11:59.132398+00:00"]