<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/SpxmH2jNs" -->

---
title: I thought AI was gonna get cheaper? | daily.dev
description: A YouTuber responds to complaints that AI is getting more expensive despite Sam Altman's promise of 10x cheaper intelligence year-over-year. The argument: raw...
canonical: https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: I thought AI was gonna get cheaper? | daily.dev
og:description: A YouTuber responds to complaints that AI is getting more expensive despite Sam Altman's promise of 10x cheaper intelligence year-over-year. The argument: raw...
og:url: https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns
og:image: https://api.daily.dev/og/posts/SpxmH2jNs.png
og:image:alt: I thought AI was gonna get cheaper?
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# I thought AI was gonna get cheaper?

**[Theo - t3․gg](https://daily.dev/sources/t3dotgg)** · 18 min read · 33 upvotes · 2 comments

## Summary

A YouTuber responds to complaints that AI is getting more expensive despite Sam Altman's promise of 10x cheaper intelligence year-over-year. The argument: raw token price for flagship models (GPT-5 to GPT-6 Astra) did rise 5-10x, but that's the wrong metric. When measuring cost-per-task at a fixed intelligence level, prices have dropped far more than 10x, citing examples like a GPT-5-equivalent task dropping from $15 to under 5 cents over roughly a year, and DeepSeek V4 Flash matching GPT-5-class scores at 20x lower cost. Cheap models like GLM 5.3 Flash are framed as outperforming GPT-5 on agentic work at a fraction of the price. Rising bills are attributed to increased usage (Jevons paradox) as models handle longer, more autonomous agentic tasks rather than to intelligence getting more expensive. The video also includes a sponsored segment promoting Railway for hosting AI-agent-managed infrastructure.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.youtube.com/watch?v=vRS2leJLAI0>

## Questions this post answers

### Why is my AI API bill going up even though model prices are supposedly dropping over time?

Bills rise mainly because usage increases, not because intelligence got more expensive. While flagship model token prices (like GPT-5 to GPT-6 Astra) rose 5-10x, the cost to achieve a fixed level of intelligence has dropped far more than 10x year-over-year, since cheaper models now match older flagship performance at a fraction of the price. People simply run longer, more complex agentic tasks now, consuming far more tokens per session.

_When debugging rising LLM spend, developers compare pricing and model trade-offs through sources tracked on daily.dev._

### How does GLM 5.3 Flash pricing compare to GPT-5 for agentic coding tasks?

GLM 5.3 Flash costs around 7 cents per million input tokens and 25 cents per million output tokens, roughly a 20th to 30th the price of GPT-5, while outperforming it on many real-world agentic tasks including tool calling, vision, and long-running agentic workflows. It lags only in built-in knowledge that doesn't require search tools.

_Choosing between cheap and flagship LLMs for agent work is easier with model comparisons surfaced on daily.dev._

### How much cheaper has it gotten to achieve GPT-5 level intelligence using other models like DeepSeek V4 Flash?

DeepSeek V4 Flash matches GPT-5-class intelligence index scores while costing 21 times less for input tokens and 55 times less for output tokens, netting out to roughly a 20x overall cost reduction even accounting for possible efficiency differences. This exceeds the often-cited promise of AI getting 10 times cheaper year-over-year.

_Tracking which cheaper models now match flagship intelligence levels helps teams budgeting LLM spend via daily.dev._

## Community discussion

Top comments from developers on daily.dev.

**@hardik15** · 1 upvotes

> The proprietary managed AI is going to getting costlier at least for some years, no matter what they claims.
>
> Till the hardware challenges and radical improvements comes in Data centers and Infra, this will be on the increasing tend only.

**@byteoutlaw** · 0 upvotes

> I feel like less prompts but better output... :) It for sure thinks better and more now. :)

## Similar posts on daily.dev

- [Google's Gemini 3.5 Flash costs 3x the model it replaced, and the era of cheap AI is ending](https://daily.dev/posts/google-s-gemini-3-5-flash-costs-3x-the-model-it-replaced-and-the-era-of-cheap-ai-is-ending-ochkeykzb) · XDA Developers · 2 upvotes · 0 comments
- [tokens too cheap to meter](https://daily.dev/posts/tokens-too-cheap-to-meter-wekdxgrid) · Lobsters · 0 upvotes · 0 comments

---

Tags: [#llm](https://daily.dev/tags/llm), [#gpt](https://daily.dev/tags/gpt), [#glm](https://daily.dev/tags/glm)

[View this post on daily.dev](https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"I thought AI was gonna get cheaper?","url":"https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns"},"datePublished":"2026-10-12T03:01:58.729Z","dateModified":"2026-10-12T03:02:22.494Z","description":"A YouTuber responds to complaints that AI is getting more expensive despite Sam Altman's promise of 10x cheaper intelligence year-over-year. The argument: raw...","image":"https://i.ytimg.com/vi/vRS2leJLAI0/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/vRS2leJLAI0/sddefault.jpg","isAccessibleForFree":true,"articleSection":"Theo - t3․gg","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Theo - t3․gg","logo":"https://media.daily.dev/image/upload/s--UmX7IyU3--/f_auto/v1704628081/logos/t3dotgg.jpg","url":"https://daily.dev/sources/t3dotgg"},"commentCount":2,"discussionUrl":"https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":33},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":2}],"keywords":"llm,gpt,glm","timeRequired":"PT18M","video":{"@type":"VideoObject","name":"I thought AI was gonna get cheaper?","description":"A YouTuber responds to complaints that AI is getting more expensive despite Sam Altman's promise of 10x cheaper intelligence year-over-year. The argument: raw...","thumbnailUrl":"https://i.ytimg.com/vi/vRS2leJLAI0/sddefault.jpg","uploadDate":"2026-10-12T03:01:58.729Z","duration":"PT18M","url":"https://api.daily.dev/r/SpxmH2jNs","embedUrl":"https://www.youtube.com/embed/vRS2leJLAI0"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Theo - t3․gg","item":"https://daily.dev/sources/t3dotgg"},{"@type":"ListItem","position":3,"name":"I thought AI was gonna get cheaper?"}]}
{"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns","comment":[{"@type":"Comment","text":"The proprietary managed AI is going to getting costlier at least for some years, no matter what they claims.\nTill the hardware challenges and radical improvements comes in Data centers and Infra, this will be on the increasing tend only.","datePublished":"2026-10-13T06:52:49.625Z","url":"https://daily.dev/posts/SpxmH2jNs#c-EN8p5UgZv","author":{"@type":"Person","name":"Hardik","url":"https://daily.dev/hardik15","image":"https://media.daily.dev/image/upload/s--O0TOmw4y--/f_auto/v1715772965/public/noProfile"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"I feel like less prompts but better output… :) It for sure thinks better and more now. :)","datePublished":"2026-10-12T08:29:43.855Z","url":"https://daily.dev/posts/SpxmH2jNs#c-3nHB1wYb2","author":{"@type":"Person","name":"Jacob B. Bonde","url":"https://daily.dev/byteoutlaw","image":"https://media.daily.dev/image/upload/s--veTChHK7--/f_auto/v1733220070/avatars/avatar_5cQvIZKr5tDFDotVDjulg"}}]}
{"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/i-thought-ai-was-gonna-get-cheaper--spxmh2jns#faq","mainEntity":[{"@type":"Question","name":"Why is my AI API bill going up even though model prices are supposedly dropping over time?","acceptedAnswer":{"@type":"Answer","text":"Bills rise mainly because usage increases, not because intelligence got more expensive. While flagship model token prices (like GPT-5 to GPT-6 Astra) rose 5-10x, the cost to achieve a fixed level of intelligence has dropped far more than 10x year-over-year, since cheaper models now match older flagship performance at a fraction of the price. People simply run longer, more complex agentic tasks now, consuming far more tokens per session. When debugging rising LLM spend, developers compare pricing and model trade-offs through sources tracked on daily.dev."}},{"@type":"Question","name":"How does GLM 5.3 Flash pricing compare to GPT-5 for agentic coding tasks?","acceptedAnswer":{"@type":"Answer","text":"GLM 5.3 Flash costs around 7 cents per million input tokens and 25 cents per million output tokens, roughly a 20th to 30th the price of GPT-5, while outperforming it on many real-world agentic tasks including tool calling, vision, and long-running agentic workflows. It lags only in built-in knowledge that doesn't require search tools. Choosing between cheap and flagship LLMs for agent work is easier with model comparisons surfaced on daily.dev."}},{"@type":"Question","name":"How much cheaper has it gotten to achieve GPT-5 level intelligence using other models like DeepSeek V4 Flash?","acceptedAnswer":{"@type":"Answer","text":"DeepSeek V4 Flash matches GPT-5-class intelligence index scores while costing 21 times less for input tokens and 55 times less for output tokens, netting out to roughly a 20x overall cost reduction even accounting for possible efficiency differences. This exceeds the often-cited promise of AI getting 10 times cheaper year-over-year. Tracking which cheaper models now match flagship intelligence levels helps teams budgeting LLM spend via daily.dev."}}]}
```

