--- title: DeepSeek-V4-Flash enters public beta with stronger agent... description: DeepSeek has launched V4-Flash as a public beta API, positioning it as a more cost-effective option for agentic workloads. Despite being a smaller model,... canonical: https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56 twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: DeepSeek-V4-Flash enters public beta with stronger agent benchmarks than V4-Pro | daily.dev og:description: DeepSeek has launched V4-Flash as a public beta API, positioning it as a more cost-effective option for agentic workloads. Despite being a smaller model,... og:url: https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56 og:image: https://api.daily.dev/og/posts/hvM60qh56.png og:image:alt: DeepSeek-V4-Flash enters public beta with stronger agent benchmarks than V4-Pro 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. # DeepSeek-V4-Flash enters public beta with stronger agent benchmarks than V4-Pro **[Collections](https://daily.dev/sources/collections)** · 2 min read · 2 upvotes · 0 comments ## Summary DeepSeek has launched V4-Flash as a public beta API, positioning it as a more cost-effective option for agentic workloads. Despite being a smaller model, V4-Flash outperforms the larger V4-Pro-Preview on agent benchmarks including Terminal Bench 2.1 (82.7), NL2Repo, and DeepSWE. It natively supports the Responses API and Codex integration. DeepSeek is also deprecating generic model identifiers like `deepseek-chat` and `deepseek-reasoner` in favor of explicit versioned names. ## Content DeepSeek's V4-Flash just went live as a public beta API, and the headline isn't the release — it's the benchmark claim: 82.7 on Terminal Bench 2.1, which puts it ahead of the larger V4-Pro-Preview on agent tasks. That's the part worth sitting with. The whole point of having a "Pro" tier is that you route the hard, expensive work there. If the Flash model is genuinely outperforming it on agentic benchmarks, that calculus breaks down. DeepSeek is saying as much directly: "We've massively upgraded its Agent capabilities — benchmark scores are now far surpassing the V4-Pro-Preview." For anyone running long agent loops, this matters more than it might look. Agent workflows are where costs compound — each step in a loop is another API call, and routing those to a heavier model adds up fast. A cheaper Flash model that actually performs better on those tasks isn't just a nice-to-have; it changes what's worth building. The reaction so far is cautious interest rather than full buy-in. Terminal Bench 2.1 is a real benchmark, but benchmark claims from the model's own developer always come with an asterisk. The community will want independent evals before anyone starts rerouting production traffic. Still, the direction is clear: DeepSeek keeps pushing the price-performance envelope in exactly the spot that hurts competitors most. Flash-tier models doing Pro-tier agent work is the kind of move that forces everyone else to either match it or explain why their pricing still makes sense. --- Tags: [#llm](https://daily.dev/tags/llm), [#ai-agents](https://daily.dev/tags/ai-agents), [#deepseek](https://daily.dev/tags/deepseek) [View this post on daily.dev](https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56) ```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":"DeepSeek-V4-Flash enters public beta with stronger agent benchmarks than V4-Pro","url":"https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56"},"datePublished":"2026-07-31T08:40:02.148Z","dateModified":"2026-07-31T09:13:52.437Z","description":"DeepSeek has launched V4-Flash as a public beta API, positioning it as a more cost-effective option for agentic workloads. Despite being a smaller model,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5d88e1fd461de15ce567540822c861bc?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/5d88e1fd461de15ce567540822c861bc?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Collections","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":"Collections","logo":"https://media.daily.dev/image/upload/s--fk_6ycEi--/f_auto,q_auto/v1780996001/logos/collections?_a=BAMAMiWQ0","url":"https://daily.dev/sources/collections"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/deepseek-v4-flash-enters-public-beta-with-stronger-agent-benchmarks-than-v4-pro-hvm60qh56","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"llm,ai-agents,deepseek","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Collections","item":"https://daily.dev/sources/collections"},{"@type":"ListItem","position":3,"name":"DeepSeek-V4-Flash enters public beta with stronger agent benchmarks than V4-Pro"}]} ```