1. Models
  2. Anthropic
  3. Misaligned behavior
AnthropicX · Jul 24, 2026 · official

Misaligned behavior

Anthropic's bar chart drew 518.2K views on X: 1.2x its median post and #145 of 1,162 AI chart posts we hold by views. Read Oct 13, 2026.

Misaligned behavior automated behavioral audit

XDataBar chartClaude MythosAll Anthropic modelsOpus 4.8Mythos 5Sonnet 5Opus 5
Published by Anthropic on X, Jul 24, 2026. Source: x.com
67/ 100

ViralChart score

Wide but thin

More views than the median post but a lower engagement rate: many saw it, few acted. Common for big accounts.

#290 of 1,162 AI chart posts we hold · #33 of 56 Anthropic posts. Score = views 40%, likes 30%, engagement rate 15%, save rate 15%, each as a percentile of every AI chart post we hold.

Our read

The post sits in the Wide but thin quadrant at #290 of 1,162 AI chart posts.

Its save rate of 0.018% sits at the 8th percentile among AI chart posts, 0.23x the 0.076% median.

It ranked #5 of 5 by views among the 5 posts from this account that day.

A score of 67 places it at #290 of 1,162 AI chart posts and #33 of 56 Anthropic posts.

20xthe median bar chart post

Across 241 bar chart posts we hold, the median gets 25.6K views.

bottom 8%save rate

0.018% of viewers bookmarked it to come back to. The median post gets 0.076%.

5posts from Anthropic that day

Part of a run of 5 posts on Jul 24, 2026. This one was #5 of 5 by views that day.

0.27views per follower

518.2K views from an account with 1.9M followers.

Public X numbers from our read on Oct 13, 2026. Peers: 1,162 AI chart posts we hold and 56 from Anthropic.

Where it sits

Reach against resonance

Each dot is a post. Right is more views, up is a higher engagement rate. The lines are the medians, so the four corners are the four kinds of post.

Among 1,162 AI chart posts we hold

Score #290 of 1,162

Views518.2K#145top 12%
Likes2.4K#126top 11%
Engagement rate0.49%#854bottom 27%
Save rate0.018%#1,071bottom 8%

Among 56 Anthropic posts

Score #33 of 56

Views518.2K#26top 46%
Likes2.4K#26top 46%
Engagement rate0.49%#35bottom 38%
Save rate0.018%#54bottom 4%

Growth

Our reads cover 28 hours

Views518.2K
Age at our read81 days
Views per day of life6.4Kviews divided by days since it was posted
Gained between our reads+10views over 28 hours, about 9 a day now

Dots are our reads. The dotted line from the day it was posted to our first read is a gap, not data.

Anthropic's posts over time

Every post from this account we hold, by date and views. This one is marked; the line is the running median of the last 10.

News · Jul 24, 2026

Anthropic posts misaligned behavior audit bar chart

The bar chart drew 518K views but a 0.018% save rate, 0.23x the AI chart median.

What happened

Anthropic published a bar chart on July 24, 2026 covering misaligned behavior automated behavioral audit. It was one of 5 posts from this account that day, alongside posts on AI Model Benchmark Comparison and Novel problem-solving by cost.

The numbers

It reached 518K views, ranking #145 of 1,162 AI chart posts against a 39.9K median. Its score of 67 ranked #290 of 1,162 AI chart posts and #33 of 56 Anthropic posts.

The lesson

A save rate of 0.018% at the 8th percentile shows this bar chart was not kept: readers saved it at 0.23x the 0.076% AI chart median.

What it means

At #33 of 56 Anthropic posts, this chart sits below the account's own top tier while its 518K views rank #145 of 1,162 AI chart posts.

Written by ViralChart from public numbers read on Oct 13, 2026. Every number is checked against our data before it is shown.

Updates: what Anthropic posted next

  1. Vulnerability identification and exploitation in open-source code669.1K views
  2. Iterative improvement on a frontier-scale model63.9K views
  3. Claude Fable 5.1 Benchmark Comparison1.2M views
  4. Terminal-Bench-Science 0.1: Accuracy vs Cost483K views
  5. Indexed cost of Fable usage694.6K views
  6. Claude Opus 5.5 Benchmark Comparison1.6M views

Data and API

4 points read off this chart

Estimated from the image 97%
SeriesGroupValueRead
Opus 4.82.85printed
Mythos 52.81printed
Sonnet 53.35printed
Opus 52.3printed

Max plan · API

GET https://viral-chart.com/api/posts/f1b117948fc4.json
GET https://viral-chart.com/api/charts/f1b117948fc4.json
GET https://viral-chart.com/api/charts/f1b117948fc4.skill.md
Make one like thisOpen the post JSONChart data JSONSkill fileAPI docs

Static files in this prototype. The post JSON holds the public numbers, ranks, score, quadrant, our reads and, when checked, our copy.

API · Max

Use this chart via API

On the Max plan you can pull this chart's numbers with the provenance of each one, as JSON or CSV, with its redraw spec, or call it from an agent over MCP. Chart id f1b117948fc4.

See the Max planAPI docs
PathReturnsPlan
GET/api/v1/charts/f1b117948fc4Title, publisher, date, form, point count; with a Max key, the data linkOpen
GET/api/v1/charts/f1b117948fc4/dataEvery point with its provenance and confidence, plus the redraw spec. Add ?format=csv for CSVMax
GET/api/v1/charts/f1b117948fc4/render?view=1The draw-only view this page usesOpen
POST/mcpTool get_chart_data with this id, for Claude, Cursor and other agentsMax
curlMax
curl -s https://viral-chart.com/api/v1/charts/f1b117948fc4/data \
  -H "Authorization: Bearer $VIRALCHART_KEY"
Claude CodeMax
claude mcp add --transport http viralchart https://viral-chart.com/mcp \
  --header "Authorization: Bearer $VIRALCHART_KEY"

60 requests a minute per key; over that you get 429 with Retry-After. Up to 5 keys per account, made on your account page and shown once. Without Max, the data route answers 402 with a link to the plans. The numbers describe a chart its publisher owns, and every file carries a license note.

More charts about Claude Mythos

Claude Mythos: Opus 4.7 Benchmark Comparison
OfficialX · AnthropicApr 16, 2026
Opus 4.7 Benchmark Comparison14.1M views · 79.6K likes
Data
Claude Mythos: Vulnerability identification and exploitation in open-source code
OfficialX · AnthropicJul 24, 2026
Vulnerability identification and exploitation in open-source code669.1K views · 2.7K likes
Data
Claude Mythos: Claude model performance in research sessions
OfficialX · AnthropicJun 4, 2026
Claude model performance in research sessions416K views · 1.6K likes
Data
Claude Mythos: Sycophancy by model
OfficialX · AnthropicApr 30, 2026
Sycophancy by model63.7K views · 170 likes
Data
Claude Mythos: Misaligned behavior
OfficialBlog · AnthropicJun 30, 2026
Misaligned behavior
Claude Mythos: Firefox 147 exploit development
OfficialBlog · AnthropicJun 30, 2026
Firefox 147 exploit development

More from Anthropic

Anthropic stacked bar chart: Firefox Security Vulnerabilities4.7x
Anthropic·Mar 2026
Firefox Security Vulnerabilities
3.2M views14.8K likes0.5% eng.
Data
Anthropic line chart: US GDP AI scenarios0.4x
Anthropic·Sep 2026
US GDP AI scenarios
856.3K views1.2K likes0.2% eng.
Data
Anthropic bar chart: RL reward hacking risks1.0x
Anthropic·Sep 2026
RL reward hacking risks
806.5K views3.2K likes0.5% eng.
Data
Claude Opus 4: Opus 4.8 benchmark comparison
OfficialX · AnthropicMay 28, 2026
Opus 4.8 benchmark comparison15.5M views · 66.3K likes
Data
Claude agents and computer use: AI Model Benchmark Comparison
OfficialX · AnthropicNov 24, 2025
AI Model Benchmark Comparison7.8M views · 18.8K likes
Data
Claude Fable 5: Claude Mythos 5 and Fable 5 Benchmark Comparison
OfficialX · AnthropicJun 9, 2026
Claude Mythos 5 and Fable 5 Benchmark Comparison6.3M views · 15.4K likes
Data
Claude Fable 5: AI Model Benchmark Comparison
OfficialX · AnthropicJul 24, 2026
AI Model Benchmark Comparison5.8M views · 10.7K likes
Data
Claude Opus 4.6: Long context MRCR v2 retrieval
OfficialX · AnthropicMar 13, 2026
Long context MRCR v2 retrieval5.7M views · 24.6K likes
Data
Claude 3.7 Sonnet: Claude 4 benchmarks
OfficialX · AnthropicMay 22, 2025
Claude 4 benchmarks4.3M views · 20.5K likes
Data

Questions people ask

Where does this chart come from?

Anthropic published it on Jul 24, 2026 on X (x.com). We keep it here, with the publisher and date credited as text, so you can study it without leaving.

Can I make a chart like this?

Use the format, not the picture. Open its data page and use Make one like this: it starts from this chart's numbers and takes yours.