--- title: Kimi K3 + Grok 4.5 Built the Same Database as Claude... description: A head-to-head benchmark pitting Kimi K3 (planning) + Grok 4.5 (implementation) against Claude Opus 5 doing both phases on the same task: building an embedded... canonical: https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Kimi K3 + Grok 4.5 Built the Same Database as Claude Opus 5 at 1/25th the Price | daily.dev og:description: A head-to-head benchmark pitting Kimi K3 (planning) + Grok 4.5 (implementation) against Claude Opus 5 doing both phases on the same task: building an embedded... og:url: https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv og:image: https://api.daily.dev/og/posts/CRIuaQyNv.png og:image:alt: Kimi K3 + Grok 4.5 Built the Same Database as Claude Opus 5 at 1/25th the Price 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. # Kimi K3 + Grok 4.5 Built the Same Database as Claude Opus 5 at 1/25th the Price **[Kilo Blog](https://daily.dev/sources/kilo-ai-blog)** · 12 min read · 1 upvotes · 0 comments ## Summary A head-to-head benchmark pitting Kimi K3 (planning) + Grok 4.5 (implementation) against Claude Opus 5 doing both phases on the same task: building an embedded key-value store in Go with strict durability, crash recovery, and atomic batch requirements. Both setups passed 64 of 65 conformance checks and shared the same critical bug. Claude Opus 5 scored 98/100 vs 93/100 for the budget combo, but the combo ran at 4% of the cost ($1.27 vs $31.71) and 23% of the time. The score gap came entirely from test depth, documentation quality, and code hygiene — not from crash safety or protocol correctness. A key finding: both planners independently introduced the same protocol bug, which neither model's own test suite caught, only an external spec-based harness did. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What critical bug did both Claude Opus 5 and the Kimi K3 plus Grok 4.5 combo produce when building the same database spec? Both implementations failed to drain a rejected oversized batch's pending operations from the TCP connection, so the server read those leftover operations as fresh standalone commands and executed them anyway. Claude Opus 5's server applied all 1,001 rejected operations, while Grok 4.5's applied 2 before the connection closed. Both planners independently wrote plans that omitted correct handling of the pending operations, so both implementers built the same flaw. _Anyone stress-testing AI-generated code for edge cases can track findings like this on daily.dev._ ### How much cheaper is using Kimi K3 for planning and Grok 4.5 for implementation compared to using Claude Opus 5 alone for a coding task? The Kimi K3 plus Grok 4.5 combo cost $1.27 versus $31.71 for Claude Opus 5, about 4% of the cost, while scoring 93 out of 100 versus 98 out of 100 on the same database-building test with 65 conformance checks. The combo also finished in 23% of the time Claude Opus 5 took, with the score gap coming entirely from test coverage, documentation, and code hygiene rather than correctness. _Teams weighing model cost against code quality can follow comparisons like this on daily.dev._ ### Why did Claude Opus 5 take much longer than Grok 4.5 to implement the same coding task? Claude Opus 5 ran 150 implementation steps over 76 minutes, with 79 shell commands in a loop of building, testing its own code, and fixing what the tests caught, producing 14 crash-scenario tests and a full-scale benchmark. Grok 4.5 finished in 11 minutes across 22 steps in essentially one shot, writing code, tests, and a README without an iterative self-review loop, resulting in a smaller test suite covering only happy paths plus one crash test. _Developers picking an AI coding agent workflow can compare these iteration styles on daily.dev._ ## Similar posts on daily.dev - [Benchmarking Kimi K3, Opus 5, Grok 4.5, and Gemini 3.6 Flash on Baba Is You](https://daily.dev/posts/benchmarking-kimi-k3-opus-5-grok-4-5-and-gemini-3-6-flash-on-baba-is-you-kjvrdfm73) · Quesma · 1 upvotes · 0 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#golang](https://daily.dev/tags/golang), [#ai-coding](https://daily.dev/tags/ai-coding) [View this post on daily.dev](https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv) ```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":"Kimi K3 + Grok 4.5 Built the Same Database as Claude Opus 5 at 1/25th the Price","url":"https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv"},"datePublished":"2026-07-30T13:07:39.550Z","dateModified":"2026-09-14T07:52:10.709Z","description":"A head-to-head benchmark pitting Kimi K3 (planning) + Grok 4.5 (implementation) against Claude Opus 5 doing both phases on the same task: building an embedded...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/933f2eb40777cd799fd3421c9765057f?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/933f2eb40777cd799fd3421c9765057f?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Kilo Blog","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":"Kilo Blog","logo":"https://media.daily.dev/image/upload/s--x3pmrf8D--/f_auto,q_auto/v1774964071/logos/kilo-ai-blog?_a=BAMAMiWQ0","url":"https://daily.dev/sources/kilo-ai-blog"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"testing,golang,ai-coding","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Kilo Blog","item":"https://daily.dev/sources/kilo-ai-blog"},{"@type":"ListItem","position":3,"name":"Kimi K3 + Grok 4.5 Built the Same Database as Claude Opus 5 at 1/25th the Price"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/kimi-k3-grok-4-5-built-the-same-database-as-claude-opus-5-at-1-25th-the-price-criuaqynv#faq","mainEntity":[{"@type":"Question","name":"What critical bug did both Claude Opus 5 and the Kimi K3 plus Grok 4.5 combo produce when building the same database spec?","acceptedAnswer":{"@type":"Answer","text":"Both implementations failed to drain a rejected oversized batch's pending operations from the TCP connection, so the server read those leftover operations as fresh standalone commands and executed them anyway. Claude Opus 5's server applied all 1,001 rejected operations, while Grok 4.5's applied 2 before the connection closed. Both planners independently wrote plans that omitted correct handling of the pending operations, so both implementers built the same flaw. Anyone stress-testing AI-generated code for edge cases can track findings like this on daily.dev."}},{"@type":"Question","name":"How much cheaper is using Kimi K3 for planning and Grok 4.5 for implementation compared to using Claude Opus 5 alone for a coding task?","acceptedAnswer":{"@type":"Answer","text":"The Kimi K3 plus Grok 4.5 combo cost $1.27 versus $31.71 for Claude Opus 5, about 4% of the cost, while scoring 93 out of 100 versus 98 out of 100 on the same database-building test with 65 conformance checks. The combo also finished in 23% of the time Claude Opus 5 took, with the score gap coming entirely from test coverage, documentation, and code hygiene rather than correctness. Teams weighing model cost against code quality can follow comparisons like this on daily.dev."}},{"@type":"Question","name":"Why did Claude Opus 5 take much longer than Grok 4.5 to implement the same coding task?","acceptedAnswer":{"@type":"Answer","text":"Claude Opus 5 ran 150 implementation steps over 76 minutes, with 79 shell commands in a loop of building, testing its own code, and fixing what the tests caught, producing 14 crash-scenario tests and a full-scale benchmark. Grok 4.5 finished in 11 minutes across 22 steps in essentially one shot, writing code, tests, and a README without an iterative self-review loop, resulting in a smaller test suite covering only happy paths plus one crash test. Developers picking an AI coding agent workflow can compare these iteration styles on daily.dev."}}]} ```