🔍 Read the full analysis: What Google’s Gemini 4 Neon Table Shows—and What It Leaves Out on ThorstenMeyerAI.com
Get business pricing on tech for your team
- 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, describing it as a model built to balance depth and efficiency. Artificial Analysis scores its highest setting one point above Gemini 4 Carbon, with lower cost and shorter task times, but its comparisons do not include the two models leading that independent index.
Google released Gemini 4 Neon on November 18, presenting it as a model that balances capability and efficiency. In Artificial Analysis measurements accessed the next day, Neon’s highest effort setting scored 61 on the Intelligence Index, one point above Gemini 4 Carbon’s high setting, while costing less and taking less time per index task. The independent results also show Neon behind the index’s two leaders, and Google’s own comparison table does not include those models.
Artificial Analysis lists Neon at $6.44 per index task and 21.5 minutes at xhigh, compared with Carbon at $7.76 and 37.1 minutes at high, using standard list prices. At Neon’s default high setting, the score is 60, matching Carbon’s cited score, for $3.74 per task and 14.8 minutes. These are benchmark measurements, not guarantees of cost or speed for every user’s workload.
Neon’s standard API price is $3 per million input tokens and $15 per million output tokens, below Carbon’s standard $4 and $20 rates. Carbon is discounted to $2 and $10 until November 30, according to the source material, so the list-price comparison does not describe the temporary offer. Artificial Analysis says Neon generates about 110 output tokens per second, versus about 55 for Carbon, but uses roughly 15% more output tokens per task, reducing some of the savings.
The model supports text, image, audio and video inputs, produces text, and has a one-million-token context window. Google’s model card gives a March 2026 knowledge cutoff and four effort settings, from low to xhigh, with high as the default. Google says Neon is adapted from Carbon; for architecture, training data and hardware, its model card refers readers to Carbon’s documentation.
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 Cost, Not a Clear Lead
The results frame Neon’s appeal less as a large jump in measured intelligence than as a cost-and-latency trade-off. For teams running long or repeated model tasks, lower benchmark task costs and faster output may matter even when the score change is small. But those figures reflect a particular index and effort setting, so buyers should test their own tasks before treating them as expected savings.
Effort selection can materially change the result. Artificial Analysis reports that moving from high to xhigh raises Neon’s index score from 60 to 61, while the cost per index task rises from $3.74 to $6.44 and time from 14.8 to 21.5 minutes. The appropriate setting may depend on whether a workload values a marginal benchmark gain more than cost and speed.
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 listed run costs are $6.86 for Neon, $15.47 for Fable and $12.61 for Opus. That is a narrow benchmark comparison, not evidence that Neon is better across software development tasks.
Products Worth Considering

LG UltraGear 27-inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDR True Black 400

Amazon Kindle Scribe Colorsoft 64GB with Premium Pen

Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite
As an affiliate, we earn on qualifying purchases.
A Fourth Gemini 4 Release
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. Google describes Carbon as suited to demanding, long-horizon projects and Hydrogen as aimed at fast everyday execution, positioning Neon between them.
The model is available through the Gemini API in Google AI Studio and Google’s business offering. Google also lists access through Antigravity and the Gemini app for paid AI Pro and Ultra subscribers. A cyber-defense version is being made available to eligible members of Google’s Fairwind Program, with access expected over the following few days, according to the supplied account.
Artificial Analysis places Neon third among model families on its Intelligence Index, behind GPT-6.2 Astra and Claude Fable 5.2. The index combines several evaluations; a ranking there does not establish that a model leads on every task. In one listed test, Terminal-Bench 4.0, Neon at xhigh scores 63.6%, below Carbon at high with 65.2%.
“frontier-level intelligence with fast end-to-end task completion”
— Google, in its Gemini 4 Neon launch post
Benchmark Gaps and Open Questions
Google’s 19-row evaluation table marks Neon best in 16 rows, but compares it with Claude Opus 5.5 and GPT-6.2 Sol, not with Claude Fable 5.2 or GPT-6.2 Astra, which lead Artificial Analysis’s index. Google’s table records rivals ahead in three rows: Opus on Terminal-Bench 4.1 and FrontierCode 2.0, and GPT-6.2 Sol on Chartography. Results across different benchmark methods should not be treated as directly interchangeable.
The supplied material does not give enough detail to assess every test’s setup, reproducibility or fit with real-world workloads. Google’s internal fabrication-rate claim is not independently verifiable from the information provided. Neon’s model card also points to Carbon documentation for architecture, training data and hardware, leaving those specifics about Neon itself unelaborated here.
Costs and timing depend on settings, task composition, token use and pricing. The cited Carbon discount expires November 30, but the material does not establish what prices will apply afterward. It also does not show whether the benchmark advantages hold across different products or customer workloads.
Real-World Tests Will Matter
The next useful evidence will come from independent evaluations that include the leading models on comparable tests, alongside published details of their methods. That would help readers judge whether Neon’s cost and speed advantages persist against the index leaders rather than only the rivals in Google’s table.
For prospective users, the immediate next step is to compare Neon’s effort settings on representative tasks and track both quality and total token costs. The high-to-xhigh comparison shows why a model name alone is not enough to predict spending or performance. The Fairwind cyber-defense rollout is also due to reach eligible members over the few days after launch, according to the supplied source; no further timing or evaluation details are provided.
Key Questions
What is Gemini 4 Neon?
Gemini 4 Neon is a Google model released November 18, 2026. Google says it is adapted from Gemini 4 Carbon and positions it between Carbon and Hydrogen in the Gemini 4 lineup.
Does Neon score much higher than Carbon?
Not in the cited Artificial Analysis comparison. Neon at xhigh scores 61, one point above Carbon at high with 60. At the high setting, Neon also scores 60.
Is Neon cheaper than Carbon?
At standard list prices and in the cited index task comparison, Neon costs less. Artificial Analysis reports $3.74 per task at high and $6.44 at xhigh. Carbon has a temporary discount to $2 per million input tokens and $10 per million output tokens through November 30, according to the supplied material.
Which models lead the independent Intelligence Index?
Artificial Analysis ranks GPT-6.2 Astra and Claude Fable 5.2 ahead of Neon among model families. Google’s cited comparison table does not include those two models.
Can users verify Google’s fabrication-rate claim?
The supplied information describes Google’s claim as based on internal Antigravity testing and says outsiders cannot check that measurement. It is distinct from the independent AA-Omniscience result cited in the source.
Source: ThorstenMeyerAI.com
