TL;DR
Google released Gemini 4 Neon on November 18, 2026, describing it as a model for fast task completion. Artificial Analysis data show a one-point Intelligence Index gain over Carbon at the tested settings, alongside lower per-task cost and shorter completion time. Google’s launch comparison omits GPT-6.2 Astra and Claude Fable 5.2, the two models ahead of Neon in that independent ranking.
Artificial Analysis reported that an index task took 21.5 minutes with Neon at xhigh, compared with 37.1 minutes for Carbon at high. The estimated cost was $6.44 per task for Neon and $7.76 for Carbon, using standard prices. These comparisons 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.
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Neon’s standard token prices are $3 per million input tokens and $15 per million output tokens, compared with Carbon’s $4 and $20. Carbon is temporarily discounted to $2 and $10 per million tokens through November 30, according to the source. Artificial Analysis also places Neon third among model families on its Intelligence Index, behind GPT-6.2 Astra and Claude Fable 5.2.
Google’s table reports Neon ahead in 16 of 19 comparisons with Claude Opus 5.5 and GPT-6.2 Sol. The source says independent measurements put Neon behind the two Intelligence Index leaders on most of the same tests. On Artificial Analysis’s Coding Agent Index, Neon scored 58.6, close to Claude Fable 5.2 at 58.5 and Opus 5.5 at 58.2; the listed task costs were $6.86, $15.47 and $12.61, respectively.

The Models Missing From Google’s Table
The choice of comparison models shapes what readers can conclude from a benchmark table. Google’s results show Neon against Opus 5.5 and GPT-6.2 Sol; they do not establish how Neon fares against Astra and Fable, which Artificial Analysis ranks ahead of it on the Intelligence Index.
For buyers, the reported cost and completion-time differences could matter as much as a one-point index change. But the figures depend on the task mix, model settings and pricing basis. Carbon’s temporary discount also changes the near-term price comparison, while Neon’s xhigh result and Carbon’s high result are not a like-for-like score test.
Products Worth Considering

Bose QuietComfort Wireless Noise Cancelling Headphones, Sandstone

Jackery Explorer 2000 v2 Portable Power Station, 2042Wh, 2200W

Apple MacBook Air 15-inch Laptop with M5 Chip, 16GB Memory, 512GB SSD, Midnight
As an affiliate, we earn on qualifying purchases.
Four Gemini Releases Since September
Neon is Google’s fourth Gemini 4 release in about seven weeks, following Argon on September 30, Carbon on October 28 and Hydrogen on November 12. Google describes Neon as a balance between Carbon, aimed at demanding long-horizon projects, and Hydrogen, intended for faster everyday execution.
Google’s model card says Neon is adapted from Gemini 4 Carbon and directs readers to Carbon’s documentation for architecture, training data and hardware details. The model accepts text, images, audio and video, has a one-million-token context window and a March 2026 knowledge cutoff, according to Google’s launch materials. It is available through several Google products and services; a cyber-defense version is being offered to eligible trusted defenders through the company’s Fairwind Program.
““frontier-level intelligence with fast end-to-end task completion.””
— Google, in its Gemini 4 Neon launch post
Limits of the Available Comparisons
The source material does not provide enough detail to determine whether Google and Artificial Analysis used identical prompts, settings or test conditions across every comparison. The Intelligence Index figures cited here also compare Neon at xhigh with Carbon at high; they should not be read as a controlled same-setting estimate of model improvement.
It is also unclear from the material how representative the index tasks are of typical customer workloads, or how costs vary across different task types. Google’s table does not include Astra or Fable, and the cited independent results do not settle how Neon compares with them across all uses.
Pricing and Benchmarks to Watch
Carbon’s stated discount is scheduled to end on November 30, 2026, which will change the price comparison if it is not extended. Further independent testing at matching effort settings, with published methods and broader model comparisons, could clarify how much of Neon’s reported advantage comes from its speed, pricing or model performance.
Google has not announced a date in the supplied material for additional benchmark results. Eligible Fairwind Program members were expected to receive access to Neon for cyber over the next few days after the launch; broader access details remain subject to Google’s availability terms.
Key Questions
When did Google release Gemini 4 Neon?
Google released Gemini 4 Neon on November 18, 2026.
Which models are missing from Google’s comparison table?
The table compares Neon with Claude Opus 5.5 and GPT-6.2 Sol. It does not include GPT-6.2 Astra or Claude Fable 5.2, which Artificial Analysis ranks ahead of Neon on its Intelligence Index.
How did Neon score against Gemini 4 Carbon?
Artificial Analysis reported a score of 61 for Neon at xhigh and 60 for Carbon at high. Because the effort settings differ, this is not a same-setting comparison.
Is Neon cheaper than Carbon?
At standard rates, Neon costs $3 per million input tokens and $15 per million output tokens, versus Carbon’s $4 and $20. Carbon is discounted to $2 and $10 per million tokens through November 30, according to the source.
Source: Thorsten Meyer AI
