What’s Missing From Google’s Gemini 4 Neon Comparison?
AIThis post was created with the assistance of artificial intelligence (AI).

🔍 Read the full analysis: What’s Missing From Google’s Gemini 4 Neon Comparison? on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys delivered free with Prime

  • Fast, free delivery on millions of items
  • Prime Video, Amazon Music and more included
  • Member-only deals all year
Start your free Prime trial Free trial for eligible customers · Cancel anytime
As an affiliate, we earn on qualifying purchases.

TL;DR

Google released Gemini 4 Neon on November 18, 2026, describing it as a model balancing depth and efficiency. Artificial Analysis scores Neon one point above Gemini 4 Carbon at its highest effort setting, with lower cost and shorter task times; however, Google’s comparison table does not include the two models that lead the independent index. Results vary by effort setting and benchmark, and the source comparison was accessed November 19.

Google released Gemini 4 Neon on November 18, positioning it as a more efficient option between its current Carbon and Hydrogen models. Figures from Artificial Analysis show Neon scoring one point above Carbon at the models’ respective highest tested settings, while costing less and taking less time per benchmark task; Google’s published comparison table, however, does not include the two models leading that independent index.

On Artificial Analysis’s Intelligence Index, Neon at xhigh scored 61, compared with Carbon at high, which scored 60. The comparison is not at the same effort setting: Neon’s default is high, where it scored 60. At that level, Artificial Analysis reports a $3.74 cost per index task; Neon at xhigh scored one point higher but cost $6.44 and took 21.5 minutes per task. Carbon at high cost $7.76 and took 37.1 minutes, based on standard list prices in the cited data.

Neon’s listed price is $3 per million input tokens and $15 per million output tokens, against Carbon’s standard $4 and $20. Carbon is discounted to $2 and $10 through November 30, according to the source article, so the standard-price comparison does not reflect that temporary offer. Artificial Analysis attributes Neon’s lower task cost partly to its lower token prices and faster output, while noting it uses about 15% more output tokens per task than Carbon.

Google’s evaluation document lists Neon ahead in 16 of 19 rows against Claude Opus 5.5 and GPT-6.2 Sol. Those are not the two models leading the Artificial Analysis Intelligence Index: the source says GPT-6.2 Astra and Claude Fable 5.2 rank ahead of Neon there. Google’s table shows Claude Opus 5.5 ahead on Terminal-bench 4.1 and FrontierCode 2.0, and GPT-6.2 Sol ahead on Chartography.

At a glance
reportWhen: Released November 18, 2026; benchmark f…
The developmentGoogle released Gemini 4 Neon, while independent benchmark figures show modest score gains over Carbon, lower task costs and an incomplete comparison set in Google’s published table.

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.

ArgonCarbonHydro.Neon ReleasedSep 30Oct 28Nov 12Nov 18 Index50605461 Cost/task$4.97$7.76$3.16$6.44 Min./task29.737.115.321.5 Output $$10$20$7.50$15 Carbon $2/$10 discount ends Nov 30Neon: +1 point, $1.32 less per task
  1. 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.
  2. 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.
  3. 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.
  4. 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 and Comparison Set Matter

The release’s practical appeal is less a dramatic benchmark jump than a possible reduction in cost and waiting time for some tasks. For teams running long reasoning or coding jobs, per-task price and completion time can affect how often a model is used and which effort setting is affordable. But the cited measurements are benchmark runs, not a guarantee that every workload will see the same savings.

The comparison set also affects how readers interpret Google’s results. A table showing Neon ahead in most rows against two selected competitors answers a narrower question than a comparison including the models at the top of the independent index. That omission does not establish that Google’s results are wrong; it means the published table does not, by itself, show how Neon performs against those two index leaders on the same rows and under the same test conditions.

Performance is not uniform across tests. Artificial Analysis reports Neon at xhigh scoring 63.6% on Terminal-Bench 4.0, below Carbon at high, at 65.2%. Its coding-agent index scores are close: Neon scored 58.6, Claude Fable 5.2 scored 58.5 and Claude Opus 5.5 scored 58.2. The reported costs per task were $6.86 for Neon, $15.47 for Fable and $12.61 for Opus. Those results are useful comparisons, but they remain tied to the index’s test methods and settings.

Products Worth Considering

FLEXISPOT Oversized Outdoor Rocking Chair with Thick Cushion, Grey

FLEXISPOT Oversized Outdoor Rocking Chair with Thick Cushion, Grey

As an affiliate, we earn on qualifying purchases.

Neon’s Place in Google’s Lineup

Neon is the 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 article. Google describes Carbon as suited to its hardest, long-horizon projects and Hydrogen as intended for fast everyday execution, with Neon balancing depth and efficiency.

Google’s model card says Neon is adapted from Gemini 4 Carbon. For architecture, training data and hardware details, the card directs readers to Carbon’s documentation rather than presenting those details separately. Neon accepts text, images, audio and video, produces text, and has a one-million-token context window. Its knowledge cutoff is March 2026, and the model offers four effort levels, with high as the default.

Google lists access through the Gemini API in AI Studio and its business offering, as well as Google Antigravity and the Gemini app for paid AI Pro and Ultra subscribers. A cyber-defense version is being made available to trusted defenders through Google’s Fairwind Program, with access for eligible members expected over the following few days, according to the supplied source.

“Neon brings “frontier-level intelligence with fast end-to-end task completion.””

— Google, in its launch post

Benchmark Limits and Open Questions

The available comparison does not establish how Neon performs against GPT-6.2 Astra and Claude Fable 5.2 across Google’s 19 evaluation rows. The supplied source says those models lead the Artificial Analysis index, but it does not provide a like-for-like table covering them across Google’s full set of tests.

Results also depend on effort level, benchmark version and pricing basis. Neon’s one-point Intelligence Index lead over Carbon compares Neon at xhigh with Carbon at high; at high, Neon’s score matches Carbon’s reported score. Carbon’s temporary discount changes the price comparison through November 30. The source material does not provide enough detail to independently reproduce every test or establish how benchmark results translate to a reader’s own workload.

Google’s launch post says Neon had its “lowest fabrication rate yet” in internal Google Antigravity testing. That is an internal result and is described as a different measurement from Artificial Analysis’s hallucination test, so it cannot be directly verified or compared using the figures supplied here. On Artificial Analysis’s AA-Omniscience test, Neon’s reported hallucination rate was 14.4%, compared with Carbon’s 11.4%.

Testing Neon Beyond the Launch

Neon is available through Google’s listed products and services, with the cyber-defense version being extended to eligible Fairwind Program members over the next few days, according to the source. The next useful evidence for buyers will be independent tests that include the omitted index leaders, use matched effort settings and disclose the benchmark and pricing conditions.

Organizations considering a switch can compare Neon’s high and xhigh settings against their own tasks, since the cited figures show that moving from high to xhigh adds one Intelligence Index point while raising the reported task cost from $3.74 to $6.44. Those figures describe Artificial Analysis’s benchmark run at standard list prices, not a forecast for any specific user. No further Google benchmark release or independent follow-up date is specified in the supplied material.

Key Questions

What changed with Gemini 4 Neon compared with Carbon?

Artificial Analysis reported Neon at xhigh scoring 61 on its Intelligence Index, one point above Carbon at high. Neon also had a lower reported cost and shorter task time in that comparison, but the models were tested at different effort settings.

Why does Google’s comparison table leave out some leading models?

Google’s table compares Neon with Claude Opus 5.5 and GPT-6.2 Sol. The supplied source says the Artificial Analysis index leaders are Claude Fable 5.2 and GPT-6.2 Astra, which do not appear in Google’s table.

Is Neon cheaper than Carbon?

At standard list prices, Neon is listed at $3 per million input tokens and $15 per million output tokens, compared with Carbon’s $4 and $20. The source says Carbon is temporarily discounted to $2 and $10 through November 30, 2026, and actual task costs vary with usage and effort setting.

Does Neon outperform Carbon on every benchmark?

No. In the supplied Artificial Analysis figures, Neon at xhigh scored 63.6% on Terminal-Bench 4.0, below Carbon at high, which scored 65.2%. Results vary across benchmarks and settings.

Where can users access Gemini 4 Neon?

Google lists Neon through the Gemini API in AI Studio, Gemini for business, and Google Antigravity and the Gemini app for paid AI Pro and Ultra subscribers. Its cyber-defense version is being offered to eligible trusted defenders through the Fairwind Program, according to the source.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.
You May Also Like

New Mud Springs Trail System Opens Near Moab, Expanding World-Class Mountain Biking Opportunities

A new trail system near Moab has opened, expanding world-class mountain biking options. Details are confirmed; impact on the biking community is significant.

From AI-Generated Code To A PDF Editor That Works Reliably

A developer reports six days of tests and redesigns after an AI-built PDF editor exposed save, encryption and redaction problems.

The Mechanics Of AI Quantization And Its Impact On Local LLMs

Exploring how advanced quantization techniques are transforming local large language model deployment and what this means for hardware and AI training.

AI And The Decision To Learn Coding: Evidence For Three Audiences

Evidence from 2026 suggests AI writes more code, but hiring, productivity and learning results point to skills beyond typing code.