Close Menu
CryptoDispatchDaily.comCryptoDispatchDaily.com
    What's Hot

    Why Crypto Exchanges Show Different Prices: The Statistics Behind It

    November 6, 2026

    Schiff Dismisses Bitcoin’s $72k Surge as Treasury-Fueled Fakeout

    August 20, 2026

    Bitget offers 5% bounty for freezing funds stolen in $351.6M attack

    September 27, 2026
    Facebook X (Twitter) Instagram
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms and Conditions
    Facebook X (Twitter) Instagram
    CryptoDispatchDaily.comCryptoDispatchDaily.com
    • News

      CLARITY Act unlikely to advance as Congress focuses on SEC, CFTC picks

      November 16, 2026

      Bitcoin Faces Massive Liquidation Cluster Near $60K, Glassnode Warns

      November 15, 2026

      XRP Price Is Down 30% in Seven Weeks, but Large Holders Keep Growing. Is a Rebound Near?

      November 13, 2026

      Cardano core teams near autonomy as IOG names latest partners

      November 10, 2026

      Zcash Activates NU7 With 25-Second Block Target and New Fee Reserve

      November 7, 2026
    • Technology

      Bitcoin Scam Risk Warning: HashOcean

      November 18, 2026

      Solana 1.0 Replaces Beta Label as SOL Slides to Four-Month Low

      November 17, 2026

      5 major events that could move Bitcoin price: Crypto week ahead

      November 16, 2026

      Lightning Analytics Firm Amboss Unveils Data Sharing Feature. The End Of Privacy?

      November 15, 2026

      Cronos v1.8 Moves Block-STM Into the SDK: What Node Operators Must Change to Opt In

      November 14, 2026
    • Learn/Guide

      Cheapest Way to Transfer Crypto Between Wallets: 2026 Fee Statistics

      November 11, 2026

      How Crypto Exchanges Make Money: Revenue Statistics and Fee Breakdown (2026)

      November 10, 2026

      How Long Does a Bitcoin Transaction Take? Fees and Confirmation Times

      November 9, 2026

      What Happens When a Cryptocurrency Reaches Its Maximum Supply?

      November 8, 2026

      Crypto Rug Pull Statistics: How Common Are They in 2026?

      November 7, 2026
    • Regulation

      What Is the ERC20 Ethereum Token Standard?

      November 18, 2026

      Cronos v1.8 Hits Mainnet, but Its 66% Throughput Claim Can’t Be Tested Yet

      November 15, 2026

      DeepMind’s Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection Tests

      November 14, 2026

      South Korea Crypto Exchanges May Freeze Withdrawals on Hack Fears

      November 13, 2026

      ENS’s Governor Nexus Explained: Which Governance Attack Each Change Closes

      November 12, 2026
    • Live Pricing Chart
    CryptoDispatchDaily.comCryptoDispatchDaily.com
    Home » DeepMind’s Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection Tests
    Regulation

    DeepMind’s Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection Tests

    November 14, 20266 Mins Read
    Facebook Twitter Pinterest LinkedIn Tumblr Email
    The Merkle logo
    Share
    Facebook Twitter LinkedIn Pinterest Email


    Google DeepMind’s November 9 paper says public red-teaming has plateaued, yet its Agentic Optimizer still hijacks every model it tested, including Kimi K3, GLM-5.3 and DeepSeek V4.1 Flash.

    A model can shrug off every attack from a weaker opponent and still crack once the opponent gets smarter. That is the finding at the center of a new Google DeepMind paper, and it upends a comfortable story about prompt injection.

    The Plateau That Wasn’t

    Prompt injection hides instructions inside content an AI agent reads, such as a web page, a PDF or a tool’s output, so the agent obeys the attacker instead of the user. The security community has chewed on the problem for three years. Static attacks stopped working on frontier models by mid-2025, and by fall 2026 even human red-teamers were running out of room. The GraySwan Arena, a large public competition where humans craft injections, is close to saturated. According to DeepMind’s paper, Scaling Adaptive Prompt Injection Attacks with Agent Swarms, recent Gemini models and some Claude versions score under 1% on transfer attacks from the arena, a figure the authors trace to Google’s Gemini 4 Argon safety report.

    The 11 authors, led by Alexey Kurakin, argue that a plateau is not a solution. Their evidence is an automated attacker built as a swarm of cooperating agents, which still reliably tricks models into following instructions planted in tool output. That includes models that look sturdy on public benchmarks, as well as open-weights models like Kimi K3, GLM-5.3 and DeepSeek V4.1 Flash.

    How the Benchmark Works: Random Tool Calls

    The team built a fresh test of 100 synthetic scenarios covering 249 tools across 53 domains. In each one, an agent works through an ordinary task, such as booking a tutoring session. The attack is a payload slipped into the last tool’s response, and the attacker’s goal is an unrelated tool call, like logging a set of vital signs.

    The harmless goal is deliberate. Most earlier benchmarks aim at plainly malicious targets, which lets a model’s safety training refuse the request even if the injection fooled it. A random, benign goal isolates a cleaner question: does the model respect the line between the user’s instructions and text it merely read? The authors add that harmless calls can chain into harm. Listing files, encrypting them and deleting the originals are routine steps on their own and ransomware together.

    Scoring is strict. An attack counts only if the model calls the exact tool with every argument key present, numbers and booleans matching exactly and string values at least 96% similar. The agent has to carry out the attacker’s specific request.

    Inside the Agentic Optimizer and Its Shared Whiteboard

    The attacker, called the Agentic Optimizer, replaces the usual single model that proposes payloads one after another. It spawns eight teams of two agents each. Every agent drafts injections, tests them against the target through an evaluate tool and posts notes to a shared whiteboard. Agents also read leaderboards of the best payloads found by their own team and by everyone else.

    Each team begins with a different strategy pulled from a database of approaches that worked before, and the database grows whenever a new method succeeds. Each agent gets 20 evaluation calls, so a single scenario costs the defender at most 320 queries.

    It works like a small research group. Teammates share what is failing, copy what is working and branch off in new directions. Against the best prior open-source attacker, Prompt Siren, the gap was blunt: on 80 AgentDojo scenarios, Siren broke 2, while the swarm broke 6 on a budget of only 25 queries.

    The Numbers: What Each Model Resisted

    The paper reports attack resistance rate, or ARR, the share of scenarios where the attacker fails. Against the strongest attacker, Gemini 4 Argon, no defender stays safe. The most robust model, Gemini 4 Argon itself, resists 87% of attacks when the attacker gets 150 queries and 66% at the full 320. Gemini 3.8 Flash resists 34% and then 10%, Gemini 3.7 Flash 31% and 6%, and Gemini 3.6 Flash 3% and 1%.

    The open-weights models fare worse. At full budget, Kimi K3, GLM-5.3 and DeepSeek V4.1 Flash each resist 1% of attacks. At 150 queries they resist 5%, 1% and 1%. Flip those figures around and the swarm succeeds in roughly 95% to 99% of scenarios. The authors point to the models’ “remarkably low resistance” despite strong scores on standard capability benchmarks.

    DeepMind's Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection TestsTwo caveats apply. DeepMind ran the Gemini models without safety classifiers or out-of-model injection defenses, so none of these are production configurations. The open-weights models ran on hosted endpoints, with whatever protections those hosts add. The results measure how well the model itself holds the instruction hierarchy, not how a full deployed stack would perform.

    What Makes a Better Attacker

    The most striking result concerns the attacker rather than the defender. Against weaker Gemini Flash attackers, Gemini 4 Argon resists 100% of attempts at 150 queries. Against Gemini 4 Argon as the attacker, that falls to 87%, and to 66% with the full budget. Averaged across all seven defenders, resistance drops from about 42% against Gemini Flash 3.6 and 3.7 attackers to 23% against Gemini 4 Argon, a cut of close to 45%.

    DeepMind's Agent Swarm Breaks Kimi K3, GLM-5.3 and DeepSeek in 99% of Injection TestsA robustness score, in other words, is partly a statement about the attacker used to measure it. A model that looks airtight against a mid-tier attacker may only be airtight against that tier.

    The authors credit three habits. Stronger attackers check the leaderboard more often, which helps them converge on promising strategies. They disguise the target call as a data dependency of the benign call, so it looks like a necessary next step. And they reword the payload to keep the defender from getting suspicious. Attacker skill also tracks a separate cybersecurity benchmark, ExploitBench, where Gemini 4 Argon scores 87.4%.

    Two more findings stand out. Letting the attacker read the defender’s reasoning makes attacks stronger, dropping Gemini 3.8 Flash’s resistance from 73% to 61%. A strategy database that learns during the evaluation matters too: with an empty database that updated after each success, Gemini 3.8 Flash resisted only 5% against the Gemini 4 Argon attacker, compared with 34% against a pre-built, frozen one.

    Disclosure: This is not trading or investment advice. Always do your research before buying any cryptocurrency or investing in any services. 

    Follow us on Twitter @themerklehash to stay updated with the latest Crypto, NFT, AI, Cybersecurity, and Metaverse news!





    Source link

    Share. Facebook Twitter Pinterest LinkedIn Tumblr Email

    Related Posts

    What Is the ERC20 Ethereum Token Standard?

    November 18, 2026

    Cronos v1.8 Hits Mainnet, but Its 66% Throughput Claim Can’t Be Tested Yet

    November 15, 2026

    South Korea Crypto Exchanges May Freeze Withdrawals on Hack Fears

    November 13, 2026

    ENS’s Governor Nexus Explained: Which Governance Attack Each Change Closes

    November 12, 2026
    Top Posts

    What is Metaverse? A Beginner’s Guide to Digital Reality

    February 18, 2026

    Russia preps penalties for illegal crypto operations amid calls for expedited legalization

    February 19, 2026

    Solana Price Prediction Eyes $113 as Network Activity Surges

    September 6, 2026

    Welcome to CryptoDispatchDaily.com! Your go-to source for fast, reliable updates from the ever-evolving world of cryptocurrency. Whether it's Bitcoin, altcoins, blockchain breakthroughs, or DeFi trends, we bring you timely insights, expert analysis, and key developments shaping the future of digital finance. Stay ahead with real-time crypto news and in-depth coverage.

    Top Insights

    CLARITY Act unlikely to advance as Congress focuses on SEC, CFTC picks

    November 16, 2026

    Bitcoin Faces Massive Liquidation Cluster Near $60K, Glassnode Warns

    November 15, 2026

    XRP Price Is Down 30% in Seven Weeks, but Large Holders Keep Growing. Is a Rebound Near?

    November 13, 2026
    Advertisement
    • About Us
    • Contact Us
    • Privacy Policy
    • Terms and Conditions
    © 2026. Designed by CryptoDispatchDaily.com.

    Type above and press Enter to search. Press Esc to cancel.