Why it worked
The video leverages a trending topic (AI tool growth) and a clear competitor comparison to create intrigue. The rapid user growth data presented visually, combined with a direct call to action, encourages engagement and positions Codex as a superior alternative.
Summary
The video discusses the explosive growth of Codex, an AI tool, highlighting its rapid user acquisition from 6 million to 25 million in just 7 weeks. It contrasts this with Claude's user frustrations and suggests that developers are switching to Codex due to its perceived better product and pricing.
Structure
- 1Introduction of Codex's explosive growth
- 2Comparison with Claude and its user issues
- 3Explanation of Codex's pricing and value proposition
- 4Call to action to comment for more information
Product placement
X.com (appears as a watermark in the top right corner, does not act, removing it would not change what happens in the video). Codex (acts as the subject of the video, its user growth is discussed and shown on a graph, removing it would change what happens in the video). Claude (acts as a competitor to Codex, its user growth is discussed, removing it would change what happens in the video). OpenAI (acts as a provider of models for Cursor, removing it would change what happens in the video). Cursor (acts as a platform that uses OpenAI models, its users are discussed as potentially switching to Codex, removing it would change what happens in the video).
Call to action
COMMENT “CODEX” if you want to see why developers are switching
On-screen text
Haider. haider1 X.com
Everyone is leaving Claude
1m -> 25m users in 7 months
6m -> 25m in basically 7 weeks
and now that openai is pulling its models from
cursor, i'd expect a decent wave of cursor
refugees who prefer to move
over to codex
CODEX
codex growth has been absolutely explosive
1m -> 25m users in 7 months
6m -> 25m in basically 7 weeks
and now that openai is pulling its models from
cursor, i'd expect a decent wave of cursor
refugees who prefer to move
over to codex
25
codex growth has been absolutely explosive
1m -> 25m users in 7 months
6m -> 25m in basically 7 weeks
and now that openai is pulling its models from
cursor, i'd expect a decent wave of cursor
refugees who prefer to move
over to codex
MILLION
codex growth has been absolutely explosive
1m -> 25m users in 7 months
6m -> 25m in basically 7 weeks
and now that openai is pulling its models from
cursor, i'd expect a decent wave of cursor
refugees who prefer to move
over to codex
IN
codex growth has been absolutely explosive
1m -> 25m users in 7 months
6m -> 25m in basically 7 weeks
and now that openai is pulling its models from
cursor, i'd expect a decent wave of cursor
refugees who prefer to move
over to codex
SEVEN WEEKS.
Codex Growth: 6m to 25m in ~7 Weeks
25m
20m
15m
10m
5m
1m+
Feb
Mar
Apr
May
Jun
Jul
Aug
THIS
AMAZING
HERE.
THINK
THE
LIKE
CODEX
WANTS
WIN
CLAWS
ACTUALLY
TABLE 1
Plan
Model
Hours per week/
scripting & 20x
Anthropic emails
Hours per week
expected if 3x
and 20x
multiplier claims
are true
Actual usage
multiplier
Pro
Sonnet 4
40-80
80-160
1x Pro
Max 5s
Sonnet 4
140-280
200-400
3.5x more usage
than Pro;
Max 20s
Sonnet 4
240-880
800-1600
6x more usage
than Pro;
1.7x more usage
than Max, 5s.
61. The hours of usage expected from Max 20x is four times that delivered by Max
5x. But the expected usage limits articulated by Defendant do not align with a multiple of four -
they are not even a multiple of two.
62. For example, Defendant's own statements about the actual range of hours it
delivers for Opus 4 on a weekly basis demonstrate that Defendant delivers only 1.1 to 1.6 times
the usage of Max 5x with the Max 20x subscription plan.
Subscription
Model
Hours delivered
per week
Monthly cost
Estimated price
per hour2
Pro
Sonnet 4
40-80 hours
$17-$20
$0.05-$0.11
Max 5s
Sonnet 4
140-280 hours
$50
$0.08-$0.16
Max 10s
Sonnet 4
140-280 hours
$50
$0.08-$0.16
Max 5s
Opus 4
15-35 hours
$5-$10
$0.14-$0.66
Max 20s
Opus 4
24-40 hours
$16-$193
THIS
WITH
200
PLANS,
REALISING
CODEX
PRODUCT.
COMMENT "CODEX"
if you want to see why
developers are switching
CODEX
if you want to see why
developers are switching
Original caption
COMMENT “CODEX” if you want to see why developers are switching 👇 Codex just hit 25 MILLION users. What’s crazy is the acceleration: it went from roughly 6M → 25M users in about 7 weeks. Meanwhile, Claude is dealing with growing frustration around usage limits, expensive Max plans, and now allegations that its advertised usage multipliers didn’t match what customers actually received. That creates a perfect opening for Codex. If developers already want GPT models — and Codex keeps getting cheaper, faster, and easier to use — we could be watching a massive shift in the AI coding war. Are you using Claude Code or Codex right now? [Codex, Claude Code, OpenAI, Anthropic, AI coding, vibe coding, coding agents, GPT, software development]#greenscreen