🔍 Read the full analysis: Gemini 4 Neon’s Small Gain And Lower Bill: Reading Between Google’s Rows 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 November 18, positioning it as a balance between capability and efficiency. Artificial Analysis scores it one point above Gemini 4 Carbon at the tested settings, while reporting lower cost and shorter completion time; results vary by effort level and benchmark.
Google released Gemini 4 Neon on November 18, presenting it as a model that balances capability and efficiency. In figures published by Artificial Analysis, Neon scores one point above Gemini 4 Carbon on the Intelligence Index at the compared settings, while its estimated cost per index task and completion time are lower; the results are benchmark measurements, not a guarantee of performance on every workload.
At its xhigh effort setting, Neon scores 61 on the Artificial Analysis Intelligence Index, compared with Carbon at high effort, which scores 60. Artificial Analysis estimates a Neon index task costs $6.44 and takes 21.5 minutes, against $7.76 and 37.1 minutes for Carbon at the standard list price. At the same high setting, Neon scores 60 for an estimated $3.74 per task. These figures compare different effort settings in the xhigh-versus-high comparison, so they should not be read as a controlled same-setting quality comparison.
Artificial Analysis attributes the cost and speed difference in part to Neon’s lower token prices and faster output. Google lists Neon at $3 per million input tokens and $15 per million output tokens, versus Carbon’s standard $4 and $20. Carbon is discounted to $2 and $10 through November 30, according to the supplied source material. Neon also uses about 15% more output tokens per index task, which offsets some of the price difference. Cached input is listed at $0.15 per million tokens.
The effort setting changes the trade-off substantially in Artificial Analysis’s data. Neon scores 60 at high for $3.74 per task and 14.8 minutes, while xhigh adds one point at $6.44 and 21.5 minutes. Medium scores 55 at $1.90 and 8.3 minutes; low scores 44 at $0.75 and 3.7 minutes. On the Coding Agent Index, Neon scores 58.6, close to Claude Fable 5.2 at 58.5 and Claude Opus 5.5 at 58.2; its estimated run cost is lower than either of those competitors in the reported comparison.
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
The Price of One More Index Point
The figures make Neon’s practical appeal less about a major jump in benchmark capability than about cost and time per task. For teams running many model-assisted coding or research jobs, lower estimated task costs could matter even if scores change little. But the savings depend on the model setting, actual usage and the applicable price: Carbon’s temporary discount narrows or reverses the list-price comparison for some customers before November 30.
The results also show why buyers need to compare performance by task rather than rely on a single overall score. Neon is close to the named Claude models on the Coding Agent Index, yet it does not lead every individual test. At xhigh, it scores 63.6% on Terminal-Bench 4.0, below Carbon’s 65.2% at high. A lower estimated cost does not by itself establish that Neon is the better choice for a particular workflow; reliability, output quality and the effort level required all matter.
Popular right now

Ninja Foodi 6-in-1 DualZone FlexBasket Air Fryer, 7-Qt MegaZone, Black, DZ071

Vitamix VX1 Professional-Grade Blender, 64 oz, Black

Breville Barista Express Manual Espresso Machine with Built-In Grinder, Brushed Stainless Steel
As an affiliate, we earn on qualifying purchases.
A Fourth Gemini 4 Model 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 material. Google describes Carbon as suited to long projects and Hydrogen to fast everyday execution, with Neon positioned between them. Its model card says Neon is adapted from Carbon and refers readers to Carbon’s documentation for details on architecture, training data and hardware.
Google says Neon accepts text, images, audio and video, returns text, and has a one-million-token context window. Its default maximum output is 64,000 tokens, with up to one million available through background long decode. The model’s stated knowledge cutoff is March 2026, and users can select low, medium, high or xhigh effort, with high as the default. Google lists access through its Gemini API and AI Studio, business offerings, selected paid subscriptions and the Gemini app for Pro and Ultra subscribers. A cyber-defense version is being made available to eligible members of Google’s Fairwind Program.
Google’s evaluation document reports Neon as best in 16 of 19 rows against Claude Opus 5.5 and GPT-6.2 Sol. Google’s own table also shows Opus ahead on Terminal-bench 4.1 and FrontierCode 2.0, and Sol ahead on Chartography. The comparison does not include Claude Fable 5.2 or GPT-6.2 Astra, which Artificial Analysis places ahead of Neon in its overall index ranking.
“Frontier-level intelligence with fast end-to-end task completion.”
— Google, in its launch post
Limits of the Benchmark Comparisons
The available figures do not establish how Neon performs across a broad range of real-world customer tasks. Artificial Analysis’s index uses defined tests, and its cost and time estimates are tied to those tests, selected effort settings and the prices in effect for the comparison. The supplied material does not specify enough detail to independently reproduce every result or determine how much performance varies with prompts and task mix.
Google’s published table and the independent index use different comparison sets and test results. Google’s table omits the two models that Artificial Analysis ranks above Neon overall, while some reported tests differ in version or setting. Google also says Neon had its lowest fabrication rate in internal Antigravity testing, but that measurement is not independently checkable from the information provided. The benchmarks therefore do not settle which model is best overall, or whether lower estimated costs translate into lower bills for every user.
Access, Pricing and Independent Tests
Neon is available through the Google products and subscription tiers listed in its launch materials. Google says eligible Fairwind Program members will receive access to the cyber-defense version over the next few days. Carbon’s stated promotional prices run through November 30, making the price comparison time-sensitive for customers choosing between the two models.
For buyers, the next useful check is to run representative tasks at the effort settings they expect to use, comparing quality, latency and total token costs rather than relying on the headline index score. Further independent testing may clarify performance across additional workloads and comparisons with the models not included in Google’s table. The supplied material does not give a date for such testing or for any broader update to Google’s published evaluation.
Key Questions
What is the main news about Gemini 4 Neon?
Google released Gemini 4 Neon on November 18, 2026. Artificial Analysis reports a one-point Intelligence Index gain over Carbon in the stated setting comparison, along with lower estimated task cost and shorter completion time.
Is Neon clearly more capable than Carbon?
The supplied benchmark results show a small overall score difference, not a clear advantage on every test. Neon scores 61 at xhigh against Carbon’s 60 at high, while Carbon scores higher on the cited Terminal-Bench 4.0 comparison.
How much does a Neon index task cost?
Artificial Analysis estimates $6.44 per task at xhigh, $3.74 at high, $1.90 at medium and $0.75 at low, using standard list prices. The estimates apply to its index tasks and are not a fixed price for every customer workload.
Does Google’s benchmark table compare Neon with the top-ranked models?
Not in the table described in the source material. Google compares Neon with Claude Opus 5.5 and GPT-6.2 Sol; Artificial Analysis’s overall index also places Claude Fable 5.2 and GPT-6.2 Astra ahead of Neon.
Where can users access Gemini 4 Neon?
Google lists access through the Gemini API and AI Studio, business offerings, Google Antigravity for paid AI Pro and Ultra subscribers, and the Gemini app for Pro and Ultra subscribers. Eligibility and product availability may vary by offering.
Source: ThorstenMeyerAI.com
