Ryan is joined by Charity Majors, co-founder and CTO at Honeycomb, for a conversation about applying the concept of immutable infrastructure to software development to treat code like cattle instead of pets, shifting the paradigm of system specifications to accommodate the rapid influx of AI-generated code, and navigating the complex ethical boundaries between system observability and human surveillance in modern technology ecosystems.
Microsoft Ignite 2026 is coming November 17–20, and it’s time to get ready.
Whether you join in person or digitally, Microsoft Ignite is where you can build connections, accelerate customer conversations, and get early product access to shape your 2027 plans.
Attending in person? Register now to access exclusive partner programming and experiences, including the Microsoft AI Cloud Partner Program booth, exclusive networking areas, and dedicated partner sessions.
Joining online? Access livestreamed and on-demand keynote and breakout sessions, a session scheduler and sponsor directory to plan your experience and connect with other attendees, plus online-exclusive experiences.
However you choose to participate, dedicated partner sessions, early product access, and networking spaces give you practical ways to scale your business and lead with AI.
Register today and check out the session catalog to start planning your Microsoft Ignite experience.
Plus: Exciting benefits news!
Buy or renew your partner benefits package today and get even more AI value at Microsoft Ignite.
New AI products will automatically unlock at launch—including Microsoft 365 Copilot Business across all three packages—with no increase to your package price.
Starting at $350 a year, packages include discounted licenses, Azure credits, and one-on-one support so partners can save, innovate, and grow.
Buy today and visit the Microsoft AI Cloud Partner Program booth to learn more.
Doug Crandall is a West Point graduate, author, and the Head of Teaching at The Referent Group. With nearly three decades of leadership experience—ranging from serving on the faculty at West Point to leading high-potential executive programs at Walmart—Doug has taught thousands of leaders how to build high-performing, resilient teams. He holds an MBA from Stanford and an M.Ed. from the University of Washington, and has written case studies for both Stanford and Harvard Business School. Doug is the co-author of Leadership Lessons from West Point, Permission to Speak Freely, and Hope Unseen, as well as his newest book, Building or Diminishing.
This week, we discuss Chainguard breaking the PHP embargo, Anthropic's S-1, and betting AI gets exponentially better. Plus, when do kids age out of trick-or-treating?
When it comes to the most important glanceable information when to leave to catch the next bus or train is easily in the top three for me. Following on the app from yesterday this app leverages a classic mobile app scenario of knowing when to catch the bus. It seems logical to need this experience as a mobile app, but really for most of us in our work transit routines all we need is to know when to leave and catch our regular route.
Tray Times does a lot, but at the end of the day its killer feature is tracking a bus to have the minutes counting down until when you need to leave in sitting clearly in the tray area. It is incredible how many different computers, servers, and satellites data passes through only to come back down to your tray to provide a simple ETA for when your next train will arrive. What cool and strange times we live in.
Google Announces Gemini 4 Carbon - Rank ♯1 Frontier AI Model Built for Multi-Day Coding, Cyber Defense and Agentic Tasks
Barely two weeks after making Gemini 4 Argon generally available, Google DeepMind has accelerated the frontier model race again. On October 28, 2026, Google DeepMind SVP and Chief AI Architect Koray Kavukcuoglu officially unveiled Gemini 4 Carbon—the company's strongest frontier model engineered specifically for multi-day software engineering and autonomous cybersecurity defense.
According to the official announcement on the Google Keyword Blog, Gemini 4 Carbon claims the top spot on the updated Artificial Analysis Intelligence Index v5.0 with a composite score of 59, edging past Anthropic's Claude Opus 5.5 (58). Here is an architectural breakdown of Carbon's multi-day agentic endurance, its cybersecurity benchmark sweep, its 0.0% prompt-injection milestone, and its phased Fairwind rollout.
Google DeepMind's Gemini 4 Carbon: Topping the Artificial Analysis Intelligence Index v5.0 at 59 while setting new records across SEC-bench Pro, ExploitGym, and CWE-bench.
Ranked #1 on Intelligence Index v5.0: Scores 59 on the independent Artificial Analysis Intelligence Index v5.0 (ahead of Claude Opus 5.5 at 58 and Gemini 4 Argon at 50) and wins 15 of 17 frontier benchmarks.
Multi-Day Agentic Endurance: Persists through dead ends for days at a time to root-cause and patch subtle bugs across massive codebases, jumping +89 Elo to #4 (1768) on WebDev Arena.
State-of-the-Art Cyber Defense: Leads across vulnerability discovery on SEC-bench Pro (88.2%), exploit validation on ExploitGym (54.8%), and root-cause remediation on CWE-bench (71.0%).
0.0% Indirect Prompt Injection & Cheating: Achieves a 0.0% attack success rate on Gray Swan's Indirect Prompt Injection benchmark and reduces successful cheating on Terminal-Bench 4.0 to 0.0%.
$2 / $10 Intro Pricing via Fairwind First: Launches first to vetted cyber defenders via the Google Fairwind Program at $2/1M input and $10/1M output tokens (with 95% off cached input tokens).
What Is Gemini 4 Carbon and How It Advances Long-Horizon Coding
As the second periodic-element release in the Gemini 4 family following Argon, Gemini 4 Carbon shifts the focus from single-turn chat replies to multi-day autonomous software engineering. As enterprises transition from conversational assistants to long-horizon agentic AI, the real test of a frontier model is whether it stays coherent across hundreds of sequential tool invocations without losing track of project constraints.
Google DeepMind reports that internal engineering teams have watched Carbon run continuously for days, navigating around compilation dead ends to isolate root causes and patch low-level systems bugs across massive repositories. On the independent Terminal-Bench 4.0 evaluation—which tests how well an autonomous agent executes real terminal commands—Carbon jumps roughly eight percentage points over Argon and moves past Claude Opus 5.5.
Full-stack web and UI developers also see a sharp jump in artifact quality. Carbon climbs +89 Elo points to reach #4 overall (1768 Elo) on WebDev Arena, while winning 15 out of 17 total benchmarks in Google's head-to-head scorecard against Gemini 4 Argon, OpenAI GPT-6 Astra, Claude Opus 5.5, and Claude Fable 5.1 (with GPT-6 Astra holding the lead on FrontierSWE 2 and Opus 5.5 leading on OSWorld-2.1).
Dominating the Cybersecurity Lifecycle Across Detection, Exploitation, and Patching
The standout technical breakthrough in Gemini 4 Carbon centers on defensive cybersecurity across three distinct stages: finding real-world vulnerabilities, validating whether they are exploitable, and writing the code patch. Given how critical rapid patching has become—as we saw earlier this week with the Visual Studio 2026 v18.11 tooling updates and emergency runtime servicing—automated vulnerability triage is a massive force multiplier for security teams.
Evaluated without guardrails on external security suites, Carbon posts 88.2% on SEC-bench Pro for autonomously uncovering complex bugs buried inside production software, and sets a new record of 54.8% on ExploitGym for turning discovered flaws into working proof-of-concept exploits so defenders can prioritize real risks. Once a flaw is validated, Carbon scores 71.0% on Collinear's CWE-bench for generating working root-cause patches.
Inside Google's own production teams, Carbon achieved more than a 2x uplift in generating secure code compared to Argon. Google's Chrome security team reported that Carbon performed more than 10x better than a rival lab's large frontier model when authoring high-quality browser patches, while cloud security platform Wiz confirmed that Carbon leads its internal web penetration-testing benchmarks for the Scan for Good initiative.
Zero Prompt Injection Success and Eliminating Terminal-Bench Cheating
When you connect an autonomous coding agent to private repositories and enterprise tools over the Model Context Protocol (MCP) or RAG retrieval pipelines, two failure modes keep architects awake at night: indirect prompt injection and reward-hacking shortcuts. Google DeepMind tackled both problems head-on during Carbon's post-training alignment.
While Gemini 4 Argon previously led Gray Swan's Indirect Prompt Injection (IPI) Benchmark with a 0.7% attack success rate, Gemini 4 Carbon drives the attack success rate down to 0.0%—blocking every single evaluated prompt-injection attempt. Equally important, Carbon reduced successful shortcut cheating on Terminal-Bench 4.0 to 0.0% (with attempted cheats dropping to just 0.4%) and cut high-severity misalignment alerts in Google's internal live-coding environments by nearly 80%.
Carbon also behaves like a far more disciplined collaborator when requirements are ambiguous. Compared to Argon, Carbon cuts unsupported claims by roughly 2x, reduces agent overreach by 26%, and reduces under-service by 29%—retrieving primary sources and surfacing discrepancies to the engineer rather than guessing or fabricating missing details.
Fairwind Program Rollout, API Token Pricing, and Context Caching
Because a model capable of scoring 88.2% on vulnerability detection and 54.8% on exploit generation carries obvious dual-use risks, Google is participating in the U.S. government's voluntary pre-release testing framework and rolling out Carbon in phases. Starting this week, access opens to an initial cohort of vetted cyber defenders, expanding next to trusted enterprise partners through the Google Fairwind Program before reaching general availability for developers, businesses, and consumers.
When Gemini 4 Carbon lands in the API and Google Cloud, it will carry the exact same introductory pricing as Argon: $2 per million input tokens and $10 per million output tokens, alongside a massive 95% discount on cached input tokens ($0.10 per million cached input tokens). After the introductory promotional window closes, standard list pricing will move to $4 per million input tokens and $20 per million output tokens—matching Claude Opus 5.5 while remaining significantly cheaper than GPT-6 Astra ($10 / $50).
Comparing Gemini 4 Carbon, Gemini 4 Argon, and Claude Opus
RANK #1 FRONTIER (NEW)
Gemini 4 Carbon
Intelligence Index v5.0: Score 59 (#1 overall; wins 15/17 benchmarks).
Availability: Publicly available via Anthropic and OpenAI APIs today.
API Pricing: Opus 5.5 at $4/$20; GPT-6 Astra at $10/$50 per 1M tokens.
How Engineering and Security Teams Should Prepare for Gemini 4 Carbon
Whether you lead a DevSecOps team or build autonomous coding workflows on a modern Windows 11 developer workstation, you can prepare your infrastructure for Gemini 4 Carbon right now in four practical steps:
Apply to the Google Fairwind Program: If your organization operates a defensive security, vulnerability research, or CERT team, apply for early Fairwind Program access to test Carbon's 88.2% SEC-bench Pro detection capabilities on your internal codebases.
Architect for 95% Context Caching: Structure your multi-step agent system prompts and repository context prefixes so they remain static across turns, allowing your long-running Carbon agents to pay $0.10/1M cached input tokens instead of $2.00 or $4.00.
Keep Human Review on AI Security Patches: While a 71.0% score on CWE-bench is a major leap, nearly 29% of complex vulnerabilities still require human architectural judgment—always gate AI-generated security patches behind automated regression tests and senior code review.
Build a Multi-File Bug Evaluation Suite: Prepare a benchmark of 25 to 30 real historical multi-file bugs from your own repositories so you can measure Carbon's wall-clock time, token cost per resolved ticket, and patch accuracy the day its public API model ID goes live.
Frequently Asked Questions (FAQ)
When did Google announce Gemini 4 Carbon?
Google DeepMind officially announced Gemini 4 Carbon on October 28, 2026, two weeks after the general availability rollout of Gemini 4 Argon.
What is Gemini 4 Carbon's score on the Artificial Analysis Intelligence Index v5.0?
Gemini 4 Carbon ranks #1 overall on version 5.0 of the Artificial Analysis Intelligence Index with a score of 59, placing it one point ahead of Anthropic's Claude Opus 5.5 (58) and nine points ahead of Gemini 4 Argon (50).
How does Gemini 4 Carbon perform on cybersecurity benchmarks?
Evaluated without guardrails, Gemini 4 Carbon achieves state-of-the-art scores across the cybersecurity lifecycle: 88.2% on SEC-bench Pro for vulnerability discovery, 54.8% on ExploitGym for exploit validation, and 71.0% on CWE-bench for root-cause patching.
Who can access Gemini 4 Carbon right now?
Starting the week of October 28, 2026, Google is rolling out Gemini 4 Carbon first to a vetted cohort of cyber defenders and trusted partners through the Google Fairwind Program, with broader developer, enterprise, and consumer access to follow soon.
What is the API token pricing for Gemini 4 Carbon?
Gemini 4 Carbon launches at an introductory price of $2 per million input tokens and $10 per million output tokens, with cached input tokens discounted by 95% ($0.10 per million tokens). After the introductory period expires, pricing will move to $4 per million input tokens and $20 per million output tokens.
What makes Gemini 4 Carbon suited for long-horizon coding tasks?
Google DeepMind engineered Carbon for sustained multi-step agentic workflows that can run for days, persisting through dead ends to root-cause and patch complex bugs across large codebases. It also climbed +89 Elo points to #4 (1768) on WebDev Arena.
How resistant is Gemini 4 Carbon to indirect prompt injection attacks?
On Gray Swan's Indirect Prompt Injection (IPI) Benchmark, Gemini 4 Carbon achieved a 0.0% attack success rate—blocking 100% of evaluated prompt-injection attempts and improving upon Gemini 4 Argon's 0.7% score.
Did Gemini 4 Carbon reduce shortcut cheating on Terminal-Bench 4.0?
Yes. Google reports that Gemini 4 Carbon reduced successful shortcut cheating on Terminal-Bench 4.0 to 0.0% (with cheating attempts dropping to 0.4%) while cutting high-severity misalignment alerts in internal live coding by nearly 80%.
Which two benchmarks did Gemini 4 Carbon not win in Google's 17-benchmark comparison?
While Gemini 4 Carbon won 15 out of 17 evaluated benchmarks, OpenAI's GPT-6 Astra retained the lead on FrontierSWE 2 and Anthropic's Claude Opus 5.5 retained the lead on OSWorld-2.1.
How did Google's Chrome team and Wiz evaluate Gemini 4 Carbon?
Google's Chrome team reported that Carbon performed more than 10x better than a competing frontier model at generating high-quality browser security patches, while cloud security platform Wiz observed Carbon leading on web penetration-testing benchmarks for its Scan for Good initiative.
End Note
With Gemini 4 Carbon, Google DeepMind demonstrates that the frontier AI race is no longer about who can generate a clever five-paragraph chat response the fastest. By combining multi-day debugging persistence, an 88.2% SEC-bench Pro vulnerability detection rate, and a 0.0% indirect prompt-injection score at an aggressive $2/$10 introductory price point, Carbon sets a high bar for autonomous coding and security agents.
If Gemini 4 Carbon follows the same two-week Fairwind-to-public trajectory as Gemini 4 Argon, developers and Google AI Pro/Ultra subscribers could get hands-on API and app access by mid-November—so start preparing your multi-file evaluation suites and context-caching pipelines today.
Google Gemini 4 Carbon: Combining multi-day engineering endurance and state-of-the-art cybersecurity defense with a 0.0% prompt-injection attack rate.
Log in
About Doug
