← All creators 
Sebastian Raschka
@rasbt Creator on X United States
Not known. Estimate about $1,000 for one sponsored post.
Open on X Checks out How to reach them
average impressions a post 210,461 Averaged across 12 posts, from X's own API. Read Sep 4, 2026. Latest post 333,878 impressions, Jul 18, 2026
followers 502,085 From X's own API. Read Sep 4, 2026.
posts every 7 days across 12 recent posts
last posted 2mo ago Read Sep 4, 2026.
Recent post performance
Views on their last 12 posts
Average views210K
ViewsAverage
Aug 14, 2025 to Jul 18, 2026. Oldest to newest.
View post data12 posts
| Post | Date | Views | Promotion flag |
|---|---|---|---|
| Gemma 3 270M! Great to see another awesome, small open-weight LLM for local tinkering. Here's a side-by-side comparison | 475,440 | Not flagged | |
| Couldn't resist. Here's a pure PyTorch from-scratch re-implementation of Gemma 3 270M in a Jupyter Notebook (uses about | 349,806 | Not flagged | |
| Pretty cool that they open sourced the actual full-sized production model. Here’s the Grok 2.5 architecture overview ne | 143,646 | Not flagged | |
| I’ve been working on something new: 📚 Build a Reasoning Model (From Scratch). The first chapters just went live! (The | 160,004 | Not flagged | |
| Updated & turned my Big LLM Architecture Comparison article into a narrated video lecture. The 11 LLM architecture | 199,766 | Not flagged | |
| When I started LLMs-from-scratch I just hoped it might help a few people learn. Just saw the GitHub the repo has now b | 233,865 | Not flagged | |
| From the Hierarchical Reasoning Model (HRM) to a new Tiny Recursive Model (TRM). A few months ago, the HRM made big wav | 128,461 | Not flagged | |
| Multi-Head Latent Attention 🔗 https://t.co/6az2lGjFqz https://t.co/CQ9ypl89v3 | 181,216 | Not flagged | |
| DeepSeek finally released a new model and paper. And because this DeepSeek-OCR release is a bit different from what ever | 158,114 | Not flagged | |
| With the release of the Kimi Linear LLM last week, we can definitely see that efficient, linear attention variants have | 70,085 | Not flagged | |
| Exciting big Kimi K2 Thinking release! More experts, fewer heads, and even more thinking! https://t.co/CxUpn68Jjj | 91,255 | Not flagged | |
| How can an LLM switch between low-, medium-, and high-effort reasoning? And how does an LLM learn to reason more or less | 333,878 | Not flagged |
Can you sponsor them?
Not known. Nothing on record either way. Ask them directly.
Estimate about $1,000 for one sponsored post
Ask them directly 502,085 followers x $2 per 1,000 followers as a rule of thumb (Sprout Social, 30 May 2024). Benchmark: Sprout Social, Influencer pricing ↗. This is our arithmetic on a published benchmark, not a quote from the creator.
What we checked
20 checks, 16 passed
| ✓ | Metrics came from the platform | Every audience figure here is a value X's own API returned, never estimated or inferred. The price estimate is the one number we work out ourselves, and it is labelled and shows its arithmetic. | not published |
| ✓ | Enough recent posts to judge | 12 recent posts sampled (target 12). | not published |
| ✓ | Average impressions ≥ 20,000 | 210,461 average impressions, median 170,610, across 12 posts. | required |
| ✓ | Typical post also clears 20,000 | Median impressions 170,610 and outlier-trimmed mean 198,001 both clear the floor. A sponsored post should land in that range. | integrity |
| ✓ | Tech / dev / software engineering | 6 distinct dev signals in the bio and post text: engineer, architecture, github, llm, gpt, pytorch. | required |
| ✓ | Not a gaming channel | No gaming signal in the text. X publishes no content category, so this is a text judgement. | required |
| ✓ | Not a hardware / gear channel | No hardware/gear pattern. | preference |
| ✓ | Target region (US / UK / EU / AU / NZ / SG) | United States (US), from the creator's self-entered location text. Self-reported by the creator, same as a platform country field. Shown with its confidence rather than treated as independently verified. | required |
| ✓ | Like rate ≥ 0.15% of impressions | 1.34% like rate, 0.032% comment rate, averaged per post. | integrity |
| ✓ | Average is not one viral outlier (X limit 15×) | Top post is 2.8× the median, within the normal spread for X, so the average reflects repeatable performance. | integrity |
| ✓ | Comment volume matches the view counts | Median 50 comments against 2,116 median likes: a normal ratio. | integrity |
| ✓ | impressions plausible for the followers count | 210,461 average impressions against 502,085 followers (42%). | integrity |
| ✓ | Posted recently | Last post 51 day(s) ago. | preference |
| ✓ | Publishes on a regular schedule | Median 7.1 day(s) between posts; the 12 sampled posts span 338 days. | preference |
| ✓ | Sampled posts are recent enough to compare | The 12 sampled posts span 338 days, close enough together that their impressions counts are comparable. | integrity |
| – | Same numbers from two endpoints of the same API | No post from this account appeared in both the search index and the profile feed, so there was nothing to cross-check. | not published |
| – | Verified against the platform's own published feed | X publishes no independent public feed to check the API's numbers against. YouTube is the only platform here where that second source exists. | not published |
| ✓ | An independent creator, not a company's own channel | x labels this a "Creator" account. | required |
| – | Same person confirmed on other platforms | No other public profile was linked from this account, so this record covers one platform only. | not published |
| – | Runs paid promotions (platform-declared) | Sponsorships shown for this creator are parsed from the text of their own post descriptions, so they are suspected rather than platform-confirmed. | not published |
Also on
| platform | handle | audience | views a post |
| Substack | @sebastianraschka | 209K |