Google has announced Gemini 4 Carbon, a new artificial intelligence model built for complex coding, autonomous workflows and cybersecurity tasks. The model aims to help developers solve difficult software problems and identify security vulnerabilities. Google says Carbon improves reasoning and coding performance for demanding engineering work.
The announcement adds another model to Google’s Gemini 4 family. It also highlights the company’s growing focus on AI systems that can complete complex, multi-step tasks.
Gemini 4 Carbon focuses on advanced coding
Gemini 4 Carbon targets difficult software engineering challenges. It can help developers investigate bugs, understand large codebases and develop fixes. These capabilities could support teams working on complex applications and software maintenance.
Google also reports improvements in full-stack web development and multimodal understanding. On WebDev Arena, Carbon gained 89 Elo points and reached fourth place with a score of 1,769, according to Google.
Google reports strong cybersecurity results
Cybersecurity is another major focus of Gemini 4 Carbon. The model aims to help security teams discover vulnerabilities, assess risks and develop patches.
Google reports an 88.2% score on SEC-bench Pro for vulnerability discovery. Carbon also scored 54.8% on ExploitGym and 71% on CWE-bench for vulnerability remediation.
Google says Carbon more than doubled the secure-code generation success rate of Gemini 4 Argon in its internal evaluations. These results reflect Google’s own testing and require independent verification.
AI safety remains a priority
Advanced AI models can introduce security risks alongside their benefits. Google says Carbon builds on the safety framework developed for Gemini 4 Argon.
The company highlights safeguards against misuse and prompt injection. It also reports improvements in alignment and fewer unsupported claims during internal testing.
Google plans a phased rollout to evaluate the model’s capabilities and safety protections. This approach will help the company gather feedback before expanding access.
Gemini 4 Carbon pricing and availability
Google is initially offering Carbon to a limited group of cybersecurity professionals. The company plans to expand access through its Fairwind Program.
Developers, businesses and consumers may receive wider access later. However, Google has not announced a specific date for a general release.
The introductory API price stands at $2 per million input tokens and $10 per million output tokens. Google plans to raise these rates to $4 and $20, respectively.
What the launch means for the AI industry
Gemini 4 Carbon reflects a broader shift towards AI systems that perform practical, complex tasks. Rather than simply answering questions, these models can assist with software development and security workflows.
Developers could use Carbon to accelerate debugging and code maintenance. Security teams could also explore its ability to identify vulnerabilities and suggest fixes.
However, real-world performance will determine how useful these capabilities become. Independent testing will help establish how Carbon compares with competing AI models.
As Google expands access, developers and cybersecurity professionals will have more opportunities to evaluate its strengths and limitations.
Read more : Google Gemini 4 Argon AI model
