🔍 Read the full analysis: Making Sense Of Gemini 4 Neon’s Results And Google’s Missing Details on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Google released Gemini 4 Neon on Nov. 18, describing it as a model that balances depth and efficiency. Artificial Analysis scores its highest setting one point above Gemini 4 Carbon, while reporting lower cost and shorter task times; Google’s published comparison does not include the two models leading that independent index.
Google released Gemini 4 Neon on Nov. 18, positioning it as a balance between the depth of Gemini 4 Carbon and the speed of Gemini 4 Hydrogen. Early figures from Artificial Analysis show Neon’s highest effort setting scoring one point above Carbon on the Intelligence Index, alongside lower reported cost per task and shorter completion times. Google’s own comparison table, however, leaves out the two models that lead that independent ranking.
Artificial Analysis gives Gemini 4 Neon xhigh a score of 61 on its Intelligence Index, compared with 60 for Gemini 4 Carbon at high. Neon ranks third among model families in that index, behind GPT-6.2 Astra and Claude Fable 5.2. The figures are not a direct comparison at identical effort settings: Neon’s 61 is for xhigh, while Carbon’s 60 is for high.
The same data put the cost of an index task at $6.44 for Neon xhigh, compared with $7.76 for Carbon high at Carbon’s standard list price. Average task time is reported as 21.5 minutes for Neon and 37.1 minutes for Carbon. At the default high setting, Neon scores 60 and costs $3.74 per index task, according to Artificial Analysis. These measurements are tied to the evaluator’s workload and should not be read as a guarantee for every use.
Google lists Neon at $3 per million input tokens and $15 per million output tokens, below Carbon’s standard rates of $4 and $20. Carbon is discounted to $2 and $10 until Nov. 30, according to the source material, so the comparison depends on whether users are paying promotional or standard prices. Artificial Analysis also reports that Neon generates output at about 110 tokens per second, versus about 55 for Carbon, while using roughly 15% more output tokens per task.
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
Lower Costs, Similar Index Scores
The release matters most to teams choosing a model for repeated, multi-step work, where per-task cost and turnaround time can affect operating budgets and how many tasks a system completes. The independent measurements suggest Neon may offer a practical efficiency improvement over Carbon, but the one-point index difference is modest, and the score changes with the selected effort level.
Artificial Analysis reports a clear spread across Neon’s settings: xhigh scores 61 for $6.44 per index task, while high scores 60 for $3.74. Medium scores 55 for $1.90, and low scores 44 for $0.75. That gives buyers a concrete trade-off to test: a small score increase at xhigh comes with a much higher task cost than the default high setting. Results from one benchmark workload do not establish which setting is best for a particular organization.
On the Artificial Analysis Coding Agent Index, Neon scores 58.6, close to Claude Fable 5.2 at 58.5 and Claude Opus 5.5 at 58.2. The reported run costs differ: $6.86 for Neon, $15.47 for Fable 5.2 and $12.61 for Opus 5.5. These figures may inform evaluation, but they do not by themselves show that Neon will be cheaper or equally capable across all coding tasks.
Products Worth Considering

Logitech MX Creative Console with 9 Customizable Keys, Graphite

LG 27GS60QC-B Ultragear 27-inch Curved Gaming Monitor (QHD, 180Hz, 1ms, 1000R)

Bose QuietComfort True Wireless Earbuds with Active Noise Cancellation, Bluetooth 5.3, USB-C Charging, and IPX4 Water Resistance
As an affiliate, we earn on qualifying purchases.
Google’s Rapid Gemini 4 Releases
Neon is the fourth Gemini 4 model Google has released in about seven weeks, following Argon on Sept. 30, Carbon on Oct. 28 and Hydrogen on Nov. 12, according to the source material. Google describes Carbon as suited to difficult, long-horizon projects and Hydrogen as aimed at fast everyday execution, with Neon between them.
Google’s model card says Neon is adapted from Gemini 4 Carbon. For details about its architecture, training data and hardware, the card refers readers to Carbon’s documentation. Google lists text, images, audio and video as inputs, text as output and a one-million-token context window. The stated knowledge cutoff is March 2026; high is the default among four effort settings, from low to xhigh.
Google says Neon is available through the Gemini API in AI Studio, Gemini for business, and the Gemini app for Pro and Ultra subscribers. Paid Google AI Pro and Ultra subscribers can also access it through Google Antigravity. A cyber-defense version is being offered to trusted defenders in Google’s Fairwind Program, with eligible members expected to receive access over the following days.
“frontier-level intelligence with fast end-to-end task completion”
— Google, in its Gemini 4 Neon launch post
Benchmark Comparisons Leave Gaps
Google’s evaluation methodology document contains 19 comparison rows, with Neon marked best in 16. Google’s table compares Neon with Claude Opus 5.5 and GPT-6.2 Sol, but does not include Claude Fable 5.2 or GPT-6.2 Astra, the two models ahead of Neon on the Artificial Analysis Intelligence Index. The different comparison sets make it harder for readers to assess Google’s headline results against the leaders in that independent ranking.
Google’s table does show rivals ahead in three rows: Claude Opus 5.5 on Terminal-bench 4.1 and FrontierCode 2.0, and GPT-6.2 Sol on Chartography. Separately, Artificial Analysis reports Neon xhigh below Carbon high on Terminal-Bench 4.0, at 63.6% versus 65.2%. Those tests and settings are not identical to every row in Google’s table, so the results should not be combined as if they were one standardized comparison.
Google says Neon had its “lowest fabrication rate yet” in internal Google Antigravity testing. The source material says that measurement is different from the outside AA-Omniscience test and cannot be checked by outside evaluators. Artificial Analysis reports a 14.4% hallucination rate for Neon on AA-Omniscience, versus 11.4% for Carbon; performance on that specific test does not independently verify Google’s internal claim. Further details about Neon’s architecture, training data and hardware are also not separately supplied in the model card described here.
Watch Access and Independent Testing
Readers can expect Neon to become available across Google’s listed products and services, while eligible Fairwind Program members are due to receive access to the cyber-defense version over the next few days. Google has not provided, in the material available here, a fuller public account of how Neon differs technically from Carbon beyond saying it is adapted from that model.
The next useful evidence will come from additional independent tests, especially comparisons that include GPT-6.2 Astra and Claude Fable 5.2 alongside Neon under clearly matched settings and workloads. Buyers evaluating the model can also compare high and xhigh on their own tasks, and check whether Carbon’s temporary discount changes the cost calculation before it expires on Nov. 30.
Key Questions
How much better did Gemini 4 Neon score than Carbon?
Artificial Analysis reports a score of 61 for Neon at xhigh and 60 for Carbon at high on its Intelligence Index. Because the settings differ, this is not a like-for-like effort comparison; Neon at high also scores 60.
Is Neon cheaper than Gemini 4 Carbon?
At standard list prices, Google’s stated token rates are $3 per million input tokens and $15 per million output tokens for Neon, compared with $4 and $20 for Carbon. Carbon is discounted to $2 and $10 until Nov. 30, according to the source material, and task costs vary by workload and effort setting.
Which leading models are missing from Google’s comparison table?
The table described in the source compares Neon with Claude Opus 5.5 and GPT-6.2 Sol. It does not include Claude Fable 5.2 or GPT-6.2 Astra, which rank ahead of Neon in the Artificial Analysis Intelligence Index.
Where can users access Gemini 4 Neon?
Google lists availability through the Gemini API in AI Studio, Gemini for business, and the Gemini app for Pro and Ultra subscribers. Paid Pro and Ultra subscribers can also use it through Google Antigravity.
Source: ThorstenMeyerAI.com
