Gemini 4 Neon Review: A Small Gain, Lower Cost, And Missing Context In Google’s Table
AIThis post was created with the assistance of artificial intelligence (AI).

TL;DR

Google released Gemini 4 Neon on November 18. Artificial Analysis gives it a score of 61, one point above Gemini 4 Carbon in the compared settings, while reporting a lower cost and shorter time per index task. Google’s launch table compares Neon with Claude Opus 5.5 and GPT-6.2 Sol, not the two models leading the independent index.

Google released Gemini 4 Neon on November 18, presenting it as a model that balances depth and efficiency. In Artificial Analysis’s independent Intelligence Index, Neon scored 61 at xhigh, one point above Gemini 4 Carbon at high; the same comparison reports a lower cost and shorter completion time per task, while Google’s launch table does not include the two models leading that index.

Before you orderOffer from Amazon

Get everyday essentials 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.

Artificial Analysis reports that one index task with Neon at xhigh costs $6.44 and takes 21.5 minutes. Its comparison figure for Carbon at high is $7.76 and 37.1 minutes. These results use different effort settings, so they do not isolate the effect of the model alone. At the same high setting, the source says Neon matches Carbon’s score for $3.74 per task.

Neon’s standard token rates are $3 per million input tokens and $15 per million output tokens, compared with Carbon’s standard $4 and $20. Carbon is discounted to $2 and $10 through November 30, according to the source. Artificial Analysis also reports roughly 110 output tokens per second for Neon and 55 for Carbon, alongside about 15% more output tokens per Neon task. The per-task comparison therefore reflects several factors, including token prices, output speed and token use.

Google’s table shows Neon ahead in 16 of 19 rows against Claude Opus 5.5 and GPT-6.2 Sol. The source says independent measurements place Neon behind GPT-6.2 Astra and Claude Fable 5.2 on most of the same tests, but does not provide the underlying row-by-row results here. On Artificial Analysis’s Coding Agent Index, Neon scores 58.6, close to Claude Fable 5.2 at 58.5 and Claude Opus 5.5 at 58.2. Reported cost per task is $6.86 for Neon, versus $15.47 and $12.61, respectively.

At a glance
reportWhen: Released November 18, 2026; figures acc…
The developmentGoogle released Gemini 4 Neon, a model adapted from Carbon that posts a small index score gain alongside lower task costs and faster completion.
Bar chart titled Same score, about half the cost per task. Artificial Analysis Intelligence Index score and cost per index task: Claude Fable 5.2 (Max) 65 and 26.54 dollars; GPT-6.2 Astra (Max) 62 and 4.00 dollars; Gemini 4 Neon (Xhigh) 61 and 6.44 dollars; Gemini 4 Neon (High) 60 and 3.74 dollars; Gemini 4 Carbon (High) 60 and 7.76 dollars; Claude Opus 5.5 (Max) 59 and 10.04 dollars; GPT-6.2 Sol (Max) 57 and 1.01 dollars. A highlighted note says Neon (High) matches Carbon (High) at 60 points for 3.74 instead of 7.76 dollars per task
Index score and cost per task at standard list prices. Source: Artificial Analysis, Gemini 4 Neon model page, accessed November 19, 2026; redrawn by Thorsten Meyer AI.

Lower Costs, Narrower Comparisons

For teams paying for model use, Neon’s reported cost and completion time per task may matter as much as its small index score change. The coding results also suggest a potential price advantage on that benchmark: Neon’s score is close to the two Claude models listed, while its reported task cost is lower. Those figures offer one basis for evaluation, but they do not establish that Neon will be cheaper or faster across every workload.

The comparison depends on the chosen setting, task and pricing window. Carbon’s temporary discount narrows or reverses some list-price comparisons until November 30. Google’s selected benchmark rivals also leave out the independent index leaders, so its table alone cannot show how Neon compares with the strongest-ranked models in that index. Buyers should compare models on their own tasks and include current rates in any cost estimate.

Products Worth Considering

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation

As an affiliate, we earn on qualifying purchases.

Four Gemini Releases Since September

Neon is Google’s fourth Gemini 4 model release in about seven weeks: Argon launched September 30, Carbon on October 28, Hydrogen on November 12 and Neon on November 18. Google describes Carbon as intended for demanding, long-horizon projects and Hydrogen for fast everyday execution, with Neon positioned between them. Its model card says Neon is adapted from Gemini 4 Carbon; for architecture, training data and hardware, the card refers readers to Carbon’s documentation.

Google lists support for text, images, audio and video as inputs, with text output and a one-million-token context window. The 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 available through the Gemini API in AI Studio and several paid Gemini products. A cyber-defense version is being offered to trusted defenders in Google’s Fairwind Program, with access for eligible members expected over the following few days.

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

— Google, in its Gemini 4 Neon launch post

Benchmark Gaps and Pricing Dates

The available comparison does not show Google’s results against GPT-6.2 Astra or Claude Fable 5.2, which the source identifies as the independent index leaders. The supplied material also does not include enough benchmark detail to reproduce every result or judge how closely the tests match a particular buyer’s workload.

The headline task-cost comparison uses Neon at xhigh and Carbon at high, while the same-setting comparison is reported separately. Actual costs and timing can vary with task, effort setting, token use and applicable discounts. The source gives no precise end date for the Fairwind access rollout beyond saying eligible members would receive access over the next few days.

Compare After Carbon’s Discount

Carbon’s stated $2 input and $10 output rates are discounted through November 30; after that, its standard rates are $4 and $20 per million tokens, according to the source. That date provides a near-term point for revisiting the price comparison. Google may also publish further evaluations or technical details, but the supplied material does not announce a date for additional benchmark results.

For now, organizations considering a switch can compare Neon with Carbon at the same effort setting and run their own tasks against competing models. Results should account for current token rates, output volume and completion time, rather than relying on a single index score or Google’s selected comparison table.

Key Questions

When did Google release Gemini 4 Neon?

Google released Gemini 4 Neon on November 18, 2026, according to the supplied launch information.

How did Neon score against Carbon?

Artificial Analysis reports a score of 61 for Neon at xhigh and 60 for Carbon at high. The settings differ; the source separately says Neon matches Carbon’s score at the same high setting.

Is Neon cheaper than Carbon?

Neon’s standard rates are $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 through November 30, so its temporary price is lower.

Did Google compare Neon with the index leaders?

The supplied account says Google’s launch table 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 lead the cited independent index.

What is still unknown about Neon’s results?

The supplied material does not provide full row-by-row benchmark data against the independent index leaders or establish how Neon performs on every real-world workload. Costs and timing may also vary with effort setting, task and token use.

Source: Thorsten Meyer AI

You May Also Like

How AI Will Replace Humans in Customer Service

Merging AI with customer service raises questions about the future of human involvement, hinting at potential shifts in industry dynamics.

Every Benchmark Launched 2023-2024 Has Fallen — The METR / SWE-Bench / CORE-Bench / MLE-Bench / PostTrainBench Sequence

Every major AI research benchmark launched in 2023-2024 has been saturated or is nearing saturation within months, indicating rapid progress in AI capabilities.

Openai Chatgpt Search Goes Global: Features and Implications

Just as OpenAI expands ChatGPT’s global search capabilities, intriguing implications arise that could change how we access information forever.

Will AI Replace Content Writers: The Future of Writing

Wondering about the fate of content writers in the age of AI? Explore the intricate balance between technology and creativity in writing.