frontpage
Select New/Existing Verizon Customers: 6-Months Google AI Pro Subscription
Free
$10
Verizon
VerizonDeal Details
Verizon offers New & Existing Customers on Simplicity, myPlan, or Verizon Home Internet Plans: 6-Months Google AI Pro Perk for Free.
Thanks Community Member AquaMeal9193 for sharing this deal
Note, Offer valid for new Google AI Pro perk enrollment only.
Google AI Pro Includes:
Thanks Community Member AquaMeal9193 for sharing this deal
Note, Offer valid for new Google AI Pro perk enrollment only.
Google AI Pro Includes:
- An upgraded Gemini, your powerful AI assistant that is far more capable at reasoning, analysis, and creative collaboration so you can get more done.
- Help with writing, planning, creating and more with Gemini in Google apps.
- Expanded access to image, music, and video generation models in Gemini.
- NotebookLM, a premium research and thinking partner to help you summarize, explore, and understand the information that matters.
- YouTube with fewer ads, offline, and in the background with YouTube Premium Lite.
- Personalized fitness, sleep, and health guidance with Google Health Premium.
- More event-based video history, intelligent alerts, and access to Gemini with Google Home Premium to make your home smarter and safer.
- 5 TB of cloud storage for Google Photos, Gmail, and Google Drive.
Editor's Notes
Written by slickdewmaster | Staff- Offer Terms:
- Account registration with Google required. Promo discount ends after 6 months or if perk is canceled or plan is moved to an ineligible plan during the promo period. After 6 months, perk bills at $10/month unless perk is canceled or unregistered. Must be 18 yrs or older. Cancel anytime. Additional terms apply. Offer subject to change.
Community Notes
About the Poster
Deal Details
Community Notes
About the Poster
Verizon offers New & Existing Customers on Simplicity, myPlan, or Verizon Home Internet Plans: 6-Months Google AI Pro Perk for Free.
Thanks Community Member AquaMeal9193 for sharing this deal
Note, Offer valid for new Google AI Pro perk enrollment only.
Google AI Pro Includes:
Thanks Community Member AquaMeal9193 for sharing this deal
Note, Offer valid for new Google AI Pro perk enrollment only.
Google AI Pro Includes:
- An upgraded Gemini, your powerful AI assistant that is far more capable at reasoning, analysis, and creative collaboration so you can get more done.
- Help with writing, planning, creating and more with Gemini in Google apps.
- Expanded access to image, music, and video generation models in Gemini.
- NotebookLM, a premium research and thinking partner to help you summarize, explore, and understand the information that matters.
- YouTube with fewer ads, offline, and in the background with YouTube Premium Lite.
- Personalized fitness, sleep, and health guidance with Google Health Premium.
- More event-based video history, intelligent alerts, and access to Gemini with Google Home Premium to make your home smarter and safer.
- 5 TB of cloud storage for Google Photos, Gmail, and Google Drive.
Editor's Notes
Written by slickdewmaster | Staff- Offer Terms:
- Account registration with Google required. Promo discount ends after 6 months or if perk is canceled or plan is moved to an ineligible plan during the promo period. After 6 months, perk bills at $10/month unless perk is canceled or unregistered. Must be 18 yrs or older. Cancel anytime. Additional terms apply. Offer subject to change.
Community Voting
Deal
Score
+114
Good Deal
Bad Deal
Top Comments
Community Reviews
This Verizon Google AI Pro perk is a six month free trial that turns into $10 per month afterward, with YouTube Premium Lite included only once the paid period starts. It also brings Gemini Notebook and strong web research tools, making it appealing for testing Google AI features. Enrollment can be finicky, though, with some signups showing shorter trial lengths, plan or device eligibility checks, and occasional need for Verizon chat support. Existing Google One or perk subscriptions may be prorated, refunded, or moved, but adding it can affect other Verizon or FIOS discounts.
The promo can prorate and refund unused time when enrolling the perk.
The perk reportedly includes YouTube Premium Lite once the paid service starts.
The enrollment process reportedly can be handled through Verizon chat support to obtain the full promotion.
Re-adding the perk after removing it reportedly can trigger the first six months free again.
The free period reportedly can be shorter than advertised, with some accounts receiving only three months.
The YouTube Premium Lite inclusion reportedly changes after the free period ends.
Adding the perk can reportedly cancel and refund an existing annual Google One subscription automatically.
Adding the perk can reportedly increase the next Verizon bill and ongoing charges beyond the expected $10 per month.
Common Questions (5)
AI-generated summary based on comments from the community. Updated ago
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!
74 Comments
Sign up for a Slickdeals account to remove this ad.
and through ChatGPT which replied "Gemini often has an advantage for broad, web-heavy information gathering, while ChatGPT and Claude often have advantages in turning research into highly polished, purpose-built deliverables. But the gap in research capability is much smaller than the statement implies."
Claude's response also supports my point more than yours: it says the claim is "roughly consistent with how these tools tend to get ranked," while only objecting that "much worse" may overstate it.
Gemini being good at web-heavy research does not mean it has the lowest hallucination rate. Those are different metrics. A model can be strong at information gathering and still be more prone to hallucinations.
So if you want to dispute my statement, show comparative hallucination benchmarks demonstrating that Gemini does not hallucinate more than ChatGPT and Claude. Simply quoting ChatGPT saying Gemini has a research advantage doesn't address the claim I actually made. You clearly do not understand what LLM is and used for.
Claude's response also supports my point more than yours: it says the claim is "roughly consistent with how these tools tend to get ranked," while only objecting that "much worse" may overstate it.
Gemini being good at web-heavy research does not mean it has the lowest hallucination rate. Those are different metrics. A model can be strong at information gathering and still be more prone to hallucinations.
So if you want to dispute my statement, show comparative hallucination benchmarks demonstrating that Gemini does not hallucinate more than ChatGPT and Claude. Simply quoting ChatGPT saying Gemini has a research advantage doesn't address the claim I actually made. You clearly do not understand what LLM is and used for.
"That reputation exists, but as a blanket statement it's becoming outdated.
Current benchmark results are mixed and depend heavily on what kind of "hallucination" you mean. On Google's FACTS suite, which tests factual accuracy across grounding, search, parametric knowledge, and multimodal tasks, current Gemini models actually rank extremely well: as of August 26, 2026, Gemini 3.7 Flash was #1, Gemini 3.1 Pro Preview #2, and GPT-5.5 #3. An independent factuality benchmark from Giskard likewise currently ranks Gemini 3.1 Pro above GPT-5.5 and Claude Opus 4.6 overall.
But there are benchmarks where Gemini looks much worse. Stanford's 2026 AI Index reproduces Artificial Analysis's AA-Omniscience results, where some Gemini variants hallucinate at very high rates—for example Gemini 3 Flash at 92%—while Gemini 3.1 Pro Preview is much better at 50%. Claude Opus 4.6 was 61% on that particular test. That sounds bizarre until you realize these tests deliberately ask models questions where they often should say they don't know. A model willing to attempt more answers can therefore score worse on "hallucination rate" while sometimes having greater overall usefulness.
Claude does have a particularly strong reputation for refusing to bluff. OpenAI's own cross-lab evaluation found Claude sometimes refused as much as 70% of very difficult factual questions, which reduced hallucinations but also reduced utility. OpenAI's models answered more often and therefore hallucinated more often in that stress test.
So I'd characterize the three more like this:
- Claude: tends to be relatively conservative and more willing to express uncertainty or decline to answer, which can make it feel much less hallucinatory.
- ChatGPT: generally good factual reliability, particularly when browsing/grounding is available, but it can still confidently extrapolate beyond evidence.
- Gemini: historically earned a reputation for some spectacular confident errors, and some Gemini variants still perform badly on particular hallucination tests. But the strongest current Gemini models are not consistently more hallucinatory than ChatGPT or Claude; on several contemporary factuality benchmarks they actually beat them.
So if what you meant was "In everyday experience Gemini seems more prone to confidently inventing specifics", I think that's a defensible observation and a fairly common perception. If you mean "benchmarks establish that Gemini in general hallucinates more than ChatGPT and Claude," the 2026 evidence no longer supports that broad claim."And Claude returns:
"The picture is more mixed than a clean ranking, but several 2026 benchmarks do lean that direction:
- On several comparisons, Gemini trails: Multiple sources put Claude around 3–4% hallucination, GPT around 6%, and Gemini around 6–9% on general factual-accuracy tests, with Gemini often "less likely to admit uncertainty."
- On AA-Omniscience (a benchmark that rewards saying "I don't know"), Gemini 3 Pro/3.1 Pro shows notably higher hallucination rates (50–88%) than Claude — but this benchmark also credits Gemini with higher raw accuracy, since it attempts more questions instead of declining.
- Other benchmarks flip the ranking: On Vectara's HHEM summarization test, some Gemini variants (especially Flash models) actually post lower hallucination rates than Claude or GPT — Gemini's real-time search grounding helps a lot on recent/factual queries.
So the honest takeaway: it's not a settled, universal fact — it depends heavily on which benchmark and task type you're looking at. Claude tends to lead on benchmarks that reward calibrated uncertainty (refusing rather than guessing), while Gemini often does better on grounded/current-events tasks due to search integration, and worse on open-ended factual recall where it answers confidently more often. ChatGPT lands in between on most of these.""That reputation exists, but as a blanket statement it's becoming outdated.
Current benchmark results are mixed and depend heavily on what kind of "hallucination" you mean. On Google's FACTS suite, which tests factual accuracy across grounding, search, parametric knowledge, and multimodal tasks, current Gemini models actually rank extremely well: as of August 26, 2026, Gemini 3.7 Flash was #1, Gemini 3.1 Pro Preview #2, and GPT-5.5 #3. An independent factuality benchmark from Giskard likewise currently ranks Gemini 3.1 Pro above GPT-5.5 and Claude Opus 4.6 overall.
But there are benchmarks where Gemini looks much worse. Stanford's 2026 AI Index reproduces Artificial Analysis's AA-Omniscience results, where some Gemini variants hallucinate at very high rates—for example Gemini 3 Flash at 92%—while Gemini 3.1 Pro Preview is much better at 50%. Claude Opus 4.6 was 61% on that particular test. That sounds bizarre until you realize these tests deliberately ask models questions where they often should say they don't know. A model willing to attempt more answers can therefore score worse on "hallucination rate" while sometimes having greater overall usefulness.
Claude does have a particularly strong reputation for refusing to bluff. OpenAI's own cross-lab evaluation found Claude sometimes refused as much as 70% of very difficult factual questions, which reduced hallucinations but also reduced utility. OpenAI's models answered more often and therefore hallucinated more often in that stress test.
So I'd characterize the three more like this:
- Claude: tends to be relatively conservative and more willing to express uncertainty or decline to answer, which can make it feel much less hallucinatory.
- ChatGPT: generally good factual reliability, particularly when browsing/grounding is available, but it can still confidently extrapolate beyond evidence.
- Gemini: historically earned a reputation for some spectacular confident errors, and some Gemini variants still perform badly on particular hallucination tests. But the strongest current Gemini models are not consistently more hallucinatory than ChatGPT or Claude; on several contemporary factuality benchmarks they actually beat them.
So if what you meant was "In everyday experience Gemini seems more prone to confidently inventing specifics", I think that's a defensible observation and a fairly common perception. If you mean "benchmarks establish that Gemini in general hallucinates more than ChatGPT and Claude," the 2026 evidence no longer supports that broad claim."And Claude returns:
"The picture is more mixed than a clean ranking, but several 2026 benchmarks do lean that direction:
- On several comparisons, Gemini trails: Multiple sources put Claude around 3–4% hallucination, GPT around 6%, and Gemini around 6–9% on general factual-accuracy tests, with Gemini often "less likely to admit uncertainty."
- On AA-Omniscience (a benchmark that rewards saying "I don't know"), Gemini 3 Pro/3.1 Pro shows notably higher hallucination rates (50–88%) than Claude — but this benchmark also credits Gemini with higher raw accuracy, since it attempts more questions instead of declining.
- Other benchmarks flip the ranking: On Vectara's HHEM summarization test, some Gemini variants (especially Flash models) actually post lower hallucination rates than Claude or GPT — Gemini's real-time search grounding helps a lot on recent/factual queries.
So the honest takeaway: it's not a settled, universal fact — it depends heavily on which benchmark and task type you're looking at. Claude tends to lead on benchmarks that reward calibrated uncertainty (refusing rather than guessing), while Gemini often does better on grounded/current-events tasks due to search integration, and worse on open-ended factual recall where it answers confidently more often. ChatGPT lands in between on most of these."I said Gemini is known for more hallucinations than ChatGPT and Claude. Your own quoted responses acknowledge that Gemini has a documented reputation for confident factual errors and performs substantially worse on some hallucination benchmarks, particularly when the model should admit that it doesn't know.
The fact that Gemini performs extremely well on some web-grounded or search-heavy benchmarks is irrelevant. I never said Gemini was bad at web research. Research capability and hallucination rate are different metrics.
And Claude being more conservative or refusing more questions doesn't somehow make Gemini's hallucinations disappear. If a model confidently invents an answer instead of saying "I don't know," that's exactly the behavior I was talking about.
So no, you haven't disproved my statement. You've demonstrated that the ranking varies by benchmark and task while simultaneously confirming that Gemini has had—and still has on some tests—a worse hallucination problem.
And yes, I understand perfectly well what LLMs are and what they are used for. You're the one changing the metric every time the evidence doesn't support your argument.
Sign up for a Slickdeals account to remove this ad.
I said Gemini is known for more hallucinations than ChatGPT and Claude. Your own quoted responses acknowledge that Gemini has a documented reputation for confident factual errors and performs substantially worse on some hallucination benchmarks, particularly when the model should admit that it doesn't know.
The fact that Gemini performs extremely well on some web-grounded or search-heavy benchmarks is irrelevant. I never said Gemini was bad at web research. Research capability and hallucination rate are different metrics.
And Claude being more conservative or refusing more questions doesn't somehow make Gemini's hallucinations disappear. If a model confidently invents an answer instead of saying "I don't know," that's exactly the behavior I was talking about.
So no, you haven't disproved my statement. You've demonstrated that the ranking varies by benchmark and task while simultaneously confirming that Gemini has had—and still has on some tests—a worse hallucination problem.
And yes, I understand perfectly well what LLMs are and what they are used for. You're the one changing the metric every time the evidence doesn't support your argument.
Join The Conversation
Share your experience with the Slickdeals community
Share information with the community. Please follow our Community Guidelines and be kind!