The first entry is always the pin's source. Overall confidence is a weighted average of how firmly each reference supports the start and end times used above; a reference counts half as much for every 180 days older than the newest.
Google’s official announcement and Reuters reporting are dated 28 October 2026; this dates the announcement, not a general public release.
Opened October 28, 2026. Evidence: “The model supports text, image, speech, and video input, outputs text, and has a 1M tokens context window.” [1]
Opened October 28, 2026. Evidence: “Terminal-Bench 4.0 (AA)” Opened comparison table and downloaded/visually inspected its matching launch benchmark graphic.[1]
Reuters corroborates the announcement date and restricted initial rollout.[1]
Neural Newscast's episode "Google Releases Gemini 4 Carbon and Connected Apps [Model Behavior]" (2026-10-28) says: "concludes. While it performs well on benchmarks like GraphWalks, for the typical enterprise, Argon is still the primary production model they can actually deploy today . [01:34] Nina Park: That is an important distinction, Thatcher. The cybersecurity focus for carbon seems very deliberate. Google claims this model more than doubles"[1]
Google launch comparison, 2026-10-28. Cybersecurity evaluations run without guardrails; benchmark results are not a deployment guarantee. · source
Google launch comparison, 2026-10-28. Cybersecurity evaluations run without guardrails; benchmark results are not a deployment guarantee. · source
Google launch comparison, 2026-10-28. Cybersecurity evaluations run without guardrails; benchmark results are not a deployment guarantee. · source
Google launch comparison, 2026-10-28. Cybersecurity evaluations run without guardrails; benchmark results are not a deployment guarantee. · source
| Benchmark | Gemini 4 Carbon | Gemini 4 Argon | GPT-6 Astra | Claude Opus 5.5 | Claude Fable 5.1 |
|---|---|---|---|---|---|
| AA Index v5.0 | 59 | 50 | 57 | 58 | 49 |
| Terminal-Bench 4.0 (AA) | 65.2% | 57.1% | 59.1% | 59.6% | 52% |
| SEC-bench Pro | 88.2% | 83.4% | 85.4% | — | — |
| ExploitGym | 54.8% | 42% | 42.4% | 34.5% | — |
| CWE-bench | 71% | 68% | 68% | 67% | 58% |
Google launch comparison, 2026-10-28. Cybersecurity evaluations run without guardrails; benchmark results are not a deployment guarantee. · source
Estimate from the version's published specs, benchmarks and reviews, 0-5; scores are only comparable within one criteria year. Gemini 4 Carbon (High), October 28 launch. AA Intelligence Index v5.0 59 => 5; Google comparison gives AA Terminal-Bench 4.0 65.2% => 5. AA Carbon-specific profile reports 1M context => 4, 102.5 output tokens/s => 3, and text/image/speech/video input => 5. Google introductory $2 input/$10 output per million gives 3:1 blended $4/M => 3.5; later $4/$20 gives $8/M, excluded from launch price score. AA lists later standard prices. One independent benchmark analysis verified (undated AA profile); fewer than three reviews because initial access is restricted. Google announcement and current DeepMind comparison opened; no Argon specifications inherited.[1]
It's also faster than average, however very verbose.
This pin follows the news: it shows what is known now, and here is what changed and what brought it.
Editedby @ThePinGang
Photos from readers, kept apart from the pin's sourced media. Rate what you ate or saw, and on a restaurant pin tag the dishes right on your photo.
Something missing or wrong? Say it in your own words: a link that backs this pin up, a different start or end date and why, or a fact it lacks or gets wrong. The AI checks it against this pin's sources, searches for better ones, and adds any page that backs you up. The pin's own sources still count most. A picture that shows something else, or shows it badly, is looked at too, and moved down or replaced.
Competes in agentic coding and advanced cybersecurity work, with access safeguards for security practitioners.
SourceCompetes in software engineering, cybersecurity and complex professional agent workflows.
SourceGoogle announced Gemini 4 Carbon on 28 October 2026. Access starts with selected cyber defenders through Fairwind; broader availability remains forthcoming.[1]
It targets cybersecurity and sustained engineering workflows. Introductory API pricing is planned at $2 per million input tokens and $10 per million output tokens.[1]
This follows Gemini 4 Argon’s release. Reuters independently reported the announcement.[4]