Google is looking to up the ante against Anthropic by assembling an elite “Coding Strike Team” to close the gap with its AI rival in coding performance, according to a report by The Information. Reportedly, Google co-founder Sergey Brin is involved in the initiative with the aim of pushing the tech giant towards the path of ‘AI takeoff’. The company is eventually planning to have self-improving AI models that can code their own upgrades.
Apart from Brin, Google DeepMind CTO Koray Kavukcuoglu has also been involved with the strike team, signalling its importance to Google leadership.
Sergey Brin takes charge:
Brin reportedly told DeepMind staffers in a recent memo that they must aggressively pivot to catch up on agents.
“To win the final sprint, we must urgently bridge the gap in agentic execution and turn our models into primary developers,” Brin was quoted as saying by The Information.
Google has made remarkable progress with its AI models in the last year or so, with the Gemini 3 series considered among the best overall models. However, Anthropic has slowly found a niche among coding-focused users who have latched on to its Claude AI to build new software by just using natural language prompts.
Meanwhile, Gemini models have historically lagged behind Anthropic’s Claude lineup in SWE-bench Verified scores, a benchmark that measures an AI’s ability to resolve real-world GitHub issues.
An earlier report by Business Insider had revealed that Google was working on an internal AI tool called Agent Smith, named after the popular antagonist from The Matrix. The AI tool can reportedly automate tasks like coding and documentation.
The report noted that Google’s AI tool works asynchronously in the background, allowing employees to check in and give it instructions directly from their phones without needing an active laptop.
Meanwhile, Google is also preparing to hold its I/O conference from 19–20 May, where the company is expected to unveil some major announcements on the AI front, including the latest Gemini models and perhaps a new text-to-video model. With the latest report indicating that coding is becoming a big priority for Google, we could well see a new initiative on that front. Notably, the company could even announce an upgrade for its Google Antigravity project, which takes on the likes of Claude Code and OpenAI’s Codex.
Google is not the only company troubled by Anthropic’s rise in the coding arena. Earlier this year, OpenAI also decided to focus its resources on winning the coding AI tool battle with Anthropic as it shut down its Sora video service and created a Codex super app by combining ChatGPT, Codex and Atlas browser.
