Skip to content
techworthit
techworthit

One question, one number. TechWorthIt reads the world's tech reviews and reporting, then tells you whether it's worth your money, with every source linked.

How we scoreAboutContactPrivacyTermsRSS
© 2026 TechWorthIt · We may earn a commission from links. It never changes a verdict.
LatestReviewsDealsSearchSaved 0
SectionExplainers / AI
AuthorTechWorthIt Desk
PublishedNov 19, 2026 · 3 min

What are AI thinking levels, and why do models have tiers?

Gemini 4 Neon shows how AI makers split models into tiers and let users dial reasoning up or down, trading answer quality against speed and cost.

A large rotary dial on a console panel turned toward its higher setting, flanked by a small hourglass and a short stack of coins.
Illustration
On this page
  1. Why do AI families have several tiers?
  2. What is a thinking level?
  3. How do tokens and context windows fit in?
  4. Why does this matter now?
  5. What does it mean for you?
0% read
On this page · 5 sections ▾
  1. Why do AI families have several tiers?
  2. What is a thinking level?
  3. How do tokens and context windows fit in?
  4. Why does this matter now?
  5. What does it mean for you?
TL;DR
AI companies sell several model sizes and let users set a thinking level because deeper reasoning costs more time and money. Gemini 4 Neon is a mid-tier example: Google says it nearly matches its larger Carbon model at a lower cost per task.

A model tier is a size and price class within one AI family, and a thinking level is a setting that controls how much reasoning the model does before it answers. Bigger tiers and higher thinking levels usually give better results on hard problems, but they take longer and cost more. Google's new Gemini 4 Neon, which 9to5Google reported sits between the Hydrogen and Carbon models, is a clear case of both ideas in one release.

Why do AI families have several tiers?

Running a large model is expensive. Every answer requires a lot of computation, and the largest models need the most. Smaller models answer faster and cost less, but they make more mistakes on complicated work.

So companies offer a range. In the Gemini 4 family, as 9to5Google laid it out, Hydrogen handles everyday help, Carbon is built for the most complex work, and Neon takes the middle slot for deeper analysis. The Gemini app's model picker now labels them by job: "Everyday help," "Deeper analysis" and "Code & complex work."

The idea is the same as choosing a vehicle. A small car is fine for errands, and a truck is better for a heavy load. Tiers let people pay for only as much capability as the task needs.

What is a thinking level?

Many current models reason before they reply. They produce intermediate steps, which are not always shown to the user, and then give a final answer. More steps usually mean a better answer on math, code or multi-part questions, and also more computation.

A thinking level is the dial for that. Neon offers four settings: Low, Medium, High and Extra High, according to the coverage. High is the default. As FoneArena described it, the Extra High setting is meant for demanding engineering, analytical and scientific tasks, while High balances performance and cost.

The practical effect is that one model can act like several. A quick question can run at a low setting, and a tricky one can get more reasoning.

How do tokens and context windows fit in?

AI models read and write in tokens, which are small chunks of text. Pricing is usually quoted per million tokens, with separate rates for input (what you send) and output (what the model writes).

Keep reading

Explainer / AI6d

What is a workhorse AI model, and why does Google want one?

Google's new Gemini 4 Hydrogen shows why AI makers split models into heavy thinkers and fast everyday ones, and what thinking levels and token prices mean for you.

A mechanical horse made of gears and metal plates pulls a cart of small cubes in front of a delicate crystal brain sculpture.
News / AI7d

Google launches Gemini 4 Hydrogen, replacing Gemini 3.8 Flash

The new fast model is 2.5x quicker than its predecessor, now powers Google Search globally, and keeps the $1.50/$7.50 API price while Argon and Carbon prices rise.

A sleek microchip on a circuit board with motion streaks, positioned ahead of a larger older chip in the background.

Related verdicts

Wait · 7.6
Xreal Aura
$1,279

Neon is priced at $3 per million input tokens and $15 per million output tokens, which 9to5Google noted is twice the price of Hydrogen. Digital Trends reported that cached input, meaning text the system has already processed and can reuse, costs $0.15 per million tokens. Higher thinking levels raise the bill indirectly, because reasoning steps consume extra output tokens.

The context window is how much material the model can hold in a single session. Neon's is 1 million tokens. FoneArena reported that it can take spreadsheets, PDFs, charts, audio and video in one window and turn them into structured results.

Why does this matter now?

The useful comparison is no longer just which model is smartest. It is how much intelligence a task gets per dollar and per minute. Google says Neon at its default High setting is effectively on par with Carbon on the Artificial Analysis Intelligence Index, at 32% lower cost per task, as Digital Trends relayed. Those are Google's figures, and Digital Trends cautioned that independent tests may differ.

Tiers also enable teams of models. FoneArena reported that Neon can act as an orchestrator in Google Antigravity, directing parallel Hydrogen subagents. A larger model plans, and cheaper ones carry out the steps. This keeps costs down on long jobs.

There is a downside. A reader of 9to5Google's comments pointed out that the growing list of model names, across Google, OpenAI and Anthropic, is hard to follow. A name like Neon says little about where it ranks, so the labels in the picker matter more than the model name.

What does it mean for you?

For most people, the choice is simple. Use the everyday tier for short questions and drafts, and switch up for long documents, research or multi-step work. Developers have one more lever: lowering the thinking level to save money, or raising it when accuracy matters more than speed.

Google says Neon is available in the Gemini app for AI Pro and Ultra subscribers, so the choice now sits in a dropdown rather than in a technical manual.

The news behind this
Google launches Gemini 4 Neon, a mid-tier model at $3/$15 →

Sources

  • 9to5GoogleGoogle launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI Pro↗
  • 9to5MacGoogle launches Gemini 4 Neon with Carbon quality at Flash speeds, live for AI Pro↗
  • FoneArenaGoogle rolls out Gemini 4 Neon with advanced coding, reasoning and research capabilities↗
  • Digital TrendsGoogle releases Gemini 4 Neon AI model, rivals Claude Opus 5.5 and GPT-6.2 Sol↗
  • India TodayGoogle launches Gemini 4 Neon for software workflows, claims 3x Carbon's speed↗

HOW THIS WAS MADE · Written by the TechWorthIt Desk with AI assistance from 5 published sources, linked above. Facts are cross-checked against those sources before publishing; we never claim testing we didn't do. Last updated Nov 19, 2026. Spotted an error? Tell us.

News / AI3h

Google Gemini for Mac adds computer use for paid subscribers

The feature lets Gemini click, type and fill forms in Mac apps and websites from a background window. It is US-only, English-only and limited to paid plans.

An open laptop with a blank screen where a cursor moves by itself across a simple empty form, as a robotic hand hovers above the keyboard.
News / AI1d

Google Gemini Tasks goes free in six countries including India

Google has made its agent feature, formerly Spark, free on the Gemini 4 Hydrogen model. Calls, purchases and bill analysis are available, though several features start in the US only.

A smartphone with abstract shapes on its screen stands beside an open gift box and a telephone handset, symbolising a free assistant feature.
techworthit
01 Reviews02 Explainers03 News