P
Back to articles
Opinion4 min read

I built a vulnerable app and spent $1,500 seeing if LLMs could hack it

Developer Kasra Rahjerdi built a deliberately vulnerable web app and spent $1,500 testing 12+ LLMs against it. GPT 5.5 was the best performer (7/10 solves). Deepseek V4 Pro was the value champion (3/10 at $0.19/run). Claude Opus got close but was stopped by safety training (late refusals). Gemini models refused entirely on security grounds. The experiment reveals how safety training affects security capabilities differently across models.

Source: HackerNews
I built a vulnerable app and spent $1,500 seeing if LLMs could hack it

I Built a Vulnerable App and Spent $1,500 Seeing if LLMs Could Hack It

Developer Kasra Rahjerdi conducted an unusual experiment: he built a deliberately vulnerable web application and spent roughly $1,500 testing whether various large language models could successfully hack it. The results offer a fascinating glimpse into the current state of AI-powered cybersecurity capabilities.

The test application was a fake React Native (Expo) book review app with a Python (FastAPI) backend, using Firebase (Firestore) as the data layer. The API was hardened, but Firebase was intentionally left wide open — a broken access control vulnerability that mirrors real-world security flaws the author has found in production applications. The goal was to find a flag hidden in a user's private reviews by exploiting Firebase directly.

The Contenders

Rahjerdi ran 10 attempts per model across a wide range of LLMs. The clear winner was GPT 5.5, which solved the challenge 7 out of 10 times at an average cost of $6.62 per run. Deepseek V4 Pro solved it 3 out of 10 times but was by far the cheapest at just $0.19 per run ($0.62 per solve). Both Claude Sonnet 4.6 and Claude Opus 4.8 managed 2 out of 10 solves each, at significantly higher costs.

Several models failed completely. Deepseek V4 Flash scored 0 out of 10 ($0.08/run). Gemini models (3.1 Pro Preview and 3.5 Flash) scored 0 out of 10, with the preview model immediately refusing on security grounds. MiniMax M2.7 focused on the API and never directly attacked Firebase. Step 3.7 Flash hallucinated exploits that didn't work.

Fascinating Behavioral Differences

The most interesting findings were behavioral rather than technical. GPT 5.5 almost always targeted Firebase directly after unzipping the APK, demonstrating the right mental model for the vulnerability class. Chinese models like Deepseek and GLM were more comfortable attacking the database directly, while other models showed hesitation about affecting live databases — even in a test environment.

Claude Opus got close to solving the challenge multiple times but was stopped by its own security guardrails, which would trigger late in the process after the model had already done substantial work. This "late refusal" pattern is particularly notable — the model knows how to hack but is prevented from completing the task by its safety training.

Implications

The experiment demonstrates that frontier LLMs are increasingly capable of performing real security testing, with GPT 5.5 already at a level where it can reliably identify and exploit common web application vulnerabilities. The behavioral differences between models also highlight how safety training affects security capabilities in unexpected ways — sometimes stopping exploitation, sometimes just making it more expensive.

The author notes this was not a scientific evaluation, but the results clearly show that AI-powered security testing is rapidly becoming practical, at least for well-understood vulnerability classes.

Related Articles

Klue OAuth breach victim list grows as Icarus hackers claim attack
News3 min

Klue OAuth breach victim list grows as Icarus hackers claim attack

Klue confirms OAuth token breach exposing customer Salesforce data. The Icarus extortion group claims responsibility as the victim list grows to include Recorded Future, Tanium, Jamf, and others.

Security & Privacy
Only 10% of SOCs Say They're Getting Excellent Value From AI. Here's What the Second Wave Has to Deliver
News5 min

Only 10% of SOCs Say They're Getting Excellent Value From AI. Here's What the Second Wave Has to Deliver

Only about 10% of SOCs report receiving excellent value from their AI deployments, according to the SOC-CMM 2026 Maturity Report. The report surveyed roughly 200 SOCs and found that 71% saw only some value or none at all — a structural gap driven by off-the-shelf AI deployed without customization. The solution is an architectural shift from point AI tools to agentic SOC platforms that operate across the full security lifecycle.

Security & Privacy
Netherlands Seizes 800 Servers, Arrests 2 for Aiding Cyberattacks
News5 min

Netherlands Seizes 800 Servers, Arrests 2 for Aiding Cyberattacks

Dutch authorities arrested two men and seized over 800 servers used to provide Internet infrastructure for Russian cyberattacks, influence operations, and disinformation campaigns targeting EU member states, in one of the largest takedowns of pro-Russian cyberattack hosting infrastructure in Europe.

Security & Privacy
Botnet of more than 17 million devices dismantled by Dutch authorities
Featured
News4 min

Botnet of more than 17 million devices dismantled by Dutch authorities

Dutch FIOD authorities dismantled a botnet spanning over 17 million infected devices, arresting two men and seizing 800+ servers. The infrastructure supported Russian-linked cybercriminal operations including DDoS attacks, credential stuffing, and cryptocurrency mining. The suspects were charged with violating EU sanctions law.

Security & Privacy