🔍 Read the full analysis: Gemini 4 Neon’s Results, Revisited: Small Gains And Missing Context on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Google released Gemini 4 Neon on November 18, 2026. Artificial Analysis scores its highest setting one point above Gemini 4 Carbon, while reporting lower cost and shorter completion time on an index task; results also vary by setting and benchmark. Google’s comparison table does not include the two models ranked ahead of Neon on the independent index.
Google released Gemini 4 Neon on November 18, positioning it as a balance of depth and efficiency. Artificial Analysis data available the next day put Neon at 61 on its Intelligence Index at xhigh, one point above Gemini 4 Carbon at high, while reporting a lower cost and shorter time for an index task; the comparison is not like-for-like across effort settings.
At standard list prices, Artificial Analysis estimates an index task with Neon at xhigh costs $6.44 and takes 21.5 minutes. Its Carbon comparison, at high, costs $7.76 and takes 37.1 minutes. At the same high setting, Neon scores 60, matching Carbon’s score in the source’s comparison, for $3.74 per task. These are benchmark measurements, not a guarantee of results or costs for other workloads.
The reported price per million tokens is $3 for input and $15 for output for Neon, compared with Carbon’s standard $4 and $20. Carbon was discounted to $2 and $10 through November 30, according to the source material, so the list-price comparison does not reflect that temporary offer. Artificial Analysis also reports Neon output at about 110 tokens per second versus Carbon at about 55, while Neon uses roughly 15% more output tokens per task, offsetting part of its price advantage.
Neon offers four effort settings: low, medium, high and xhigh. In Artificial Analysis’s figures, its index scores range from 44 at low to 61 at xhigh; measured task cost rises from $0.75 to $6.44, and time from 3.7 to 21.5 minutes. Google says high is the default. The model accepts text, images, audio and video, produces text, and has a one-million-token context window, according to Google’s launch post and model card.
Gemini 4 Neon
Four models in seven weeks
Google’s Gemini 4 releases side by side, each at its best tested setting in the independent Artificial Analysis index. The best value in each row is outlined.
- Four models in seven weeksArgon on September 30, Carbon on October 28, Hydrogen on November 12, Neon on November 18. Neon is adapted from Carbon.
- A small step in the indexNeon at xhigh scores 61, one point above Carbon at high. It ranks third among model families, behind GPT-6.2 Astra and Claude Fable 5.2.
- A bigger step in cost and timeOne index task costs $6.44 instead of $7.76 and takes 21.5 instead of 37.1 minutes. Neon’s tokens are 25% cheaper than Carbon’s standard price.
- What to watchUntil November 30, Carbon still sells at a discounted $2/$10. Google’s own table compares Neon with Claude Opus 5.5 and GPT-6.2 Sol, not with the index leaders.
Source: Artificial Analysis, Gemini 4 Neon model page and article, accessed Nov 19, 2026 (Intelligence Index v5.0.1, cost and time per task at standard prices, release dates); Google, “Introducing Gemini 4 Neon”, Nov 18, 2026; Gemini 4 Neon model card. Output price in US$ per million tokens, standard rates. Redrawn by Thorsten Meyer AI.
Graphic: Powered by Thorsten Meyer AI
Cost Gains Come With Setting Choices
The release puts a practical trade-off ahead of a dramatic benchmark leap: the measured score increase over Carbon is small, while cost and speed differ more in the reported index run. For teams weighing deployments, the effort setting matters: moving from high to xhigh adds one index point but raises the measured cost from $3.74 to $6.44 per task. The relevant choice will depend on whether a particular job needs the additional measured performance.
Neon also ranks third among model families on the Artificial Analysis Intelligence Index, behind GPT-6.2 Astra and Claude Fable 5.2. On the separate Coding Agent Index, the source reports scores of 58.6 for Neon, 58.5 for Fable 5.2 and 58.2 for Claude Opus 5.5—within half a point. Neon’s measured coding run cost $6.86 per task, versus $15.47 for Fable 5.2 and $12.61 for Opus 5.5. The figures offer a cost-and-performance comparison, not proof that one model will be cheaper or better for every user.
Products Worth Considering

Canon SELPHY CP1500 Wireless Photo Printer Bundle with 108 Sheets of Photo Paper

Canon SELPHY QX20 Wireless Compact Photo Printer Bundle (Dark Gray)

Dell UltraSharp U4025QW 40-Inch Curved 5K2K WUHD LED Monitor
As an affiliate, we earn on qualifying purchases.
A Fourth Gemini Release in Seven Weeks
Neon is Google’s fourth Gemini 4 model in about seven weeks, following Argon on September 30, Carbon on October 28 and Hydrogen on November 12, according to the source. Google describes Carbon as suited to difficult, long-running projects and Hydrogen to fast everyday execution, with Neon between them. Its model card says Neon is adapted from Gemini 4 Carbon; for architecture, training data and hardware, it points readers to Carbon’s documentation.
Google’s evaluation document lists 19 benchmark rows and marks Neon best in 16. In the three rows where the document shows a rival ahead, Opus 5.5 leads on Terminal-bench 4.1 and FrontierCode 2.0, while GPT-6.2 Sol leads on Chartography. The table includes Opus 5.5 and GPT-6.2 Sol, but not Fable 5.2 or GPT-6.2 Astra, the two models ranked above Neon on the independent index. Google’s table and Artificial Analysis’s index use different comparisons, so their results should not be treated as a single head-to-head ranking.
“lowest fabrication rate yet in internal Google Antigravity testing”
— Google’s launch post
Benchmarks Leave Comparisons Open
The available figures do not establish how Neon performs across every real-world task. Artificial Analysis’s central comparison pairs Neon at xhigh with Carbon at high, while its same-setting high comparison gives both a score of 60. Costs and completion times also depend on the specific benchmark, model setting and token use.
Results are not uniformly in Neon’s favor. On Terminal-Bench 4.0, Artificial Analysis reports Neon at xhigh scoring 63.6%, below Carbon at high at 65.2%. Its AA-Omniscience hallucination rate is 14.4%, compared with Carbon’s 11.4%; the source says rates for several other high-setting models were higher. Google’s claim about a low fabrication rate comes from internal Antigravity testing, which is not the same measure and cannot be independently checked from the supplied material.
Google’s published table does not include the two models ahead of Neon on the independent index. The supplied information also does not provide a full methodology comparison that would reconcile the different evaluation tables. The reported knowledge cutoff is March 2026; the source gives no further detail on how the model’s adaptation affects behavior or capability beyond the benchmark figures.
Access and Further Testing
Google says Neon is available through the Gemini API in Google AI Studio, Gemini for business, and Google Antigravity for paid AI Pro and Ultra subscribers. The Gemini app availability listed in the source is also for Pro and Ultra subscribers. Google said a cyber-defense version, Gemini 4 Neon for cyber, would reach eligible members of its Fairwind Program over the following few days; the supplied material does not confirm when that access was completed.
The next useful evidence will come from further independent evaluations, especially comparisons that include GPT-6.2 Astra and Claude Fable 5.2 under clearly matched settings and report the same task, price assumptions and time window. Users can also test high versus xhigh on their own workloads, since the benchmark results show that the extra effort setting changes both measured score and cost. No such later results or revised comparisons are included in the material available here.
Key Questions
What did Google release?
Google released Gemini 4 Neon on November 18, 2026. Its model card says it is adapted from Gemini 4 Carbon.
How much better did Neon score than Carbon?
Artificial Analysis reported an Intelligence Index score of 61 for Neon at xhigh and 60 for Carbon at high. At the same high setting, Neon also scored 60, so the one-point difference applies to the cited xhigh-versus-high comparison, not a matched-setting comparison.
Is Neon cheaper than Carbon?
Neon’s standard listed prices are $3 per million input tokens and $15 per million output tokens, below Carbon’s standard $4 and $20. The source says Carbon had a temporary discount to $2 and $10 through November 30, 2026, and benchmark-task costs vary by setting and workload.
Did Neon lead every benchmark?
No. Google’s table marks Neon best in 16 of 19 rows, while showing rivals ahead in three. Artificial Analysis also reports Neon below Carbon on Terminal-Bench 4.0 and ranks GPT-6.2 Astra and Claude Fable 5.2 ahead of Neon on its Intelligence Index.
Where can users access Gemini 4 Neon?
The source lists the Gemini API in Google AI Studio, Gemini for business, Google Antigravity for paid AI Pro and Ultra subscribers, and the Gemini app for those subscribers. It says the cyber-defense version was due to reach eligible Fairwind Program members over the next few days after launch.
Source: ThorstenMeyerAI.com
