--- title: Kimi K3 is making Opus-tier models look overpriced description: Moonshot AI's Kimi K3 has landed with benchmark results placing it third behind Fable 5 and Opus 4.8, but hands-on agentic testing suggests it outperforms both... canonical: https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Kimi K3 is making Opus-tier models look overpriced | daily.dev og:description: Moonshot AI's Kimi K3 has landed with benchmark results placing it third behind Fable 5 and Opus 4.8, but hands-on agentic testing suggests it outperforms both... og:url: https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw og:image: https://api.daily.dev/og/posts/9272HSLzW.png og:image:alt: Kimi K3 is making Opus-tier models look overpriced 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 is making Opus-tier models look overpriced **[Trends](https://daily.dev/sources/trends)** · 2 min read · 2 upvotes · 0 comments ## Summary Moonshot AI's Kimi K3 has landed with benchmark results placing it third behind Fable 5 and Opus 4.8, but hands-on agentic testing suggests it outperforms both in practical use — using Chrome CLI unprompted, staying on task through long workflows, and generally feeling more capable than its rank implies. The developer community is drawing a sharp conclusion: if open models are already trading blows with Anthropic's flagship on real agentic tasks, the pricing model for closed frontier models looks increasingly hard to justify. A possible next iteration called 'Kivine' has appeared on LMSYS Arena, suggesting K3 may not be the ceiling. The practical tip circulating is to use K3 via Kimi CLI rather than the web interface for better agentic performance. The broader takeaway: the next frontier model release needs to be a genuine leap, not an incremental bump at a premium price. ## Content Moonshot AI dropped Kimi K3 this week: 2.8 trillion parameters, 1M token context, open weights arriving July 27. The benchmarks are real. The hype is partly wrong. On the numbers, K3 scores 57 on the Artificial Analysis Intelligence Index, putting it roughly level with Opus 4.8 and just behind GPT-5.6 Sol (55) and Claude Fable 5. It hit #1 in the Frontend Code Arena with 1,679 points, beating Fable 5. On Epoch's ECI it scores 156, edging out Opus 4.6 and GPT-5.3-Codex. One hands-on test placed it third overall (77.5%) behind Fable 5 and Opus 4.8, but the reviewer said it *felt* better than both on real agentic work. That gap between benchmark rank and practical feel is where most of the interesting discourse lives. The "cheap Chinese model" framing is getting demolished in real time. K3 is priced at $3/$15 per million input/output tokens. It burns roughly 1.9x more output tokens than GPT-5.6 Sol to complete the same task. Do the math: $0.94 per task vs $1.04 for Sol. The discount evaporated. One analyst put it bluntly: "K3's massive token bloat completely swallowed up its massive discount on paper." Emad Mostaque argues inference costs will drop 10-50x as US infrastructure companies optimize the weights, but that's a future bet, not today's reality. The cybersecurity angle is getting attention too. K3 reportedly fixed 15 critical security bugs in 10 hours for $250 after Codex and Fable refused due to guardrails. Whether that's a feature or a warning sign depends entirely on who you ask. The "open weights" framing has its own asterisk. At 2.8T parameters, you need roughly 1.4TB of storage before quantization overhead. Reddit's top joke: "2TB VRAM Is All You Need." One sharp observer noted that "open weights increasingly means auditable by companies with GPU clusters, not runnable by you." The weights also dropped from the docs briefly after launch before reappearing, which didn't go unnoticed. There's also a distillation whisper going around: K3 apparently identified itself as Claude in at least one screenshot. Moonshot hasn't addressed it. The market read this as a DeepSeek repeat and sold off AI stocks. That's probably overdone. But the pattern of capable Chinese open models arriving at frontier-adjacent quality, on a faster and faster cadence, is getting harder to wave away. --- Tags: [#ai](https://daily.dev/tags/ai), [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#kimi-k3](https://daily.dev/tags/kimi-k3) [View this post on daily.dev](https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw) ```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 is making Opus-tier models look overpriced","url":"https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw"},"datePublished":"2026-07-16T13:55:19.551Z","dateModified":"2026-07-23T10:53:43.630Z","description":"Moonshot AI's Kimi K3 has landed with benchmark results placing it third behind Fable 5 and Opus 4.8, but hands-on agentic testing suggests it outperforms both...","image":"https://i.ytimg.com/vi/JnC2MuAEOiQ/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/JnC2MuAEOiQ/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Trends","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":"Trends","logo":"https://media.daily.dev/image/upload/s--ZfSp3asX--/f_auto,q_auto/v1780996004/logos/trends?_a=BAMAMiWQ0","url":"https://daily.dev/sources/trends"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/kimi-k3-is-making-opus-tier-models-look-overpriced-9272hslzw","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,llm,ai-agents,kimi-k3","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Trends","item":"https://daily.dev/sources/trends"},{"@type":"ListItem","position":3,"name":"Kimi K3 is making Opus-tier models look overpriced"}]} ```