← Back to leaderboard
iam_elias1

Elias

dormant⚠ low-confidence
@iam_elias1

Sharing my AI journey with everyone. From research papers to daily tips to building agents. Posting what resonates. DM for collab. eliaskennedy368@gmail.com

RosterAI Model News & ReleasesTech Career & Coding EducationAI Agents & Productivity Tools
35,111 followers · 608 following · Last polled 9h ago
Perf. Score
23.5
Rank #348
Value Score
27.3
#201 of 273 priced
Median views
883
p25 883 · trailing 7d
Followers
35K
-0.2% 7d
ER (impr.)
0.57%
eng ÷ impressions
ER (foll.)
0.01%
eng ÷ followers
Est. CPM
$56.63
QT-rate basis
WoW views
-34.5%
▼ falling

Rates & value · est. from 883 median organic views

overpriced
Quote tweetbasis
$50
$56.63 / 1K views
Post
—
no CPM
Thread
—
no CPM
Retweet
—
no CPM
Value Score 27.3 · rank #201 of 273 priced$10000 / 1K engagementsrate age unknown (pre-audit import)

Trends

Follower growth
Median views / day
Engagement rate
View distribution · trailing-7d posts

Each blue dot is one organic post. The Performance Score ranks by the median (883), not the mean (883) — so a single viral post doesn't inflate a creator's reach. The p25 floor (883) is the level 75% of posts clear.

Not enough in-window posts yet to show a distribution.

Recent posts · ranked by views

PostPostedViewsLikesRepostsRepliesQuotesBookmarksER
A lot of AI video creation still starts with a prompt. What caught my attention here is how much more focus Dreamina AI is putting on the workflow itself. #DreaminaCaughtTheVibe @Dreamina_ai updated the Dreamina AI web experience and canvas, and it’s live now. The new experience brings AI video workflows and skills for film, brand ads, viral social content, and more. The useful part is having a more structured way to move from an idea into something you can actually keep building on. Get 90% OFF your first month of Dreamina Basic until October 9th. I’d test one of the workflows on a real project and see how it fits into your process.frozen28d ago6.2K1255040.42%
Space Bunny reads images and codes like it has somewhere to be. Gave it a napkin with a messy app idea → it read every word and doodle, then built a working hydration tracker app ️Screenshotted the running app → it reviewed the UI and improved it without being told what was wrong Asked for a full-stack upgrade → Express server, SQLite database, REST API with 8 routes. 16 minutes. Napkin to full-stack. One session. Free on OpenCode until ~Sep 30 🐰
frozen
26d ago
5.4K
24
11
3
0
4
0.78%
I going to test this model on Opencode. For context, this model was launched by a food delivery company in China.frozen26d ago2.5K900120.48%
I have tested Longcat 2.5 preview, as it was free in Opencode. The things I noticed- >It is fast, quite fast. >In frontend it performs somewhat closer to Muse spark 1.3. >The result was on point and concise, less bullshit. >The overall experience was smooth, fast. But it is still far from being a frontier model. Even models like Deepseek V4.1 flash mogs this model completely.frozen26d ago1.9K532010.59%
Singapore just wrapped, and Los Angeles is next. CapCut World Singapore brought creators from across the CapCut ecosystem together to share their ideas, their process, and the stories behind their work. A global creative community in one room. → Creators from around the world, sharing what they make and how → AI-powered creativity front and center → New creative tools built for video creators → A community that keeps growing CapCut keeps building for the people who make things, from AI creators to video editors to anyone with an idea worth showing. Next stop: CapCut World Los Angeles, USA, October 2026. If you create, this one's for you. #CapCutWorld #CapCutDidThatfrozen9d ago1.6K18120032.01%
Opus 5.5 vs GPT-6 ASTRA. Opus is clearly the winner here. https://t.co/v93XktS2Onfrozen1mo ago1.6K810000.56%
Nobody expected anything to beat Hermes from @NousResearch as the default agent. This side-by-side is the first time I've seen it happen. OpenHuman finishes well ahead, and you can watch the context window and the running cost in real time while it works. If I'm choosing an agent after seeing this, it's OpenHuman from @tinyhumansai over Hermesfrozen17d ago1.4K742031.16%
I am with NIKfrozen18d ago1.2K400000.33%
This seems a pretty solid model but small sized. Will test it soonfrozen22d ago1.2K100000.08%
Sonnet 5.5 just launched and it is close to Opus 5.5 in benchmark. Will drop a test soonfrozen24d ago1.2K600010.59%
Just say that you cut 50% of usage for the Pro plan.frozen23d ago1.2K300000.26%
I created this website completely free. Used Space bunny inside Opencode(Free) And got this banger looking website. I didn't even write the prompts myself. Full tutorial is coming soon! https://t.co/SgaF6z9Rx9frozen22d ago1.2K16101022.52%
You gotta be kidding me. Google just dropped Gemini 4 Argon and it beat Astra, Opus 5.5? 🤯🤯frozen22d ago1.1K300000.26%
Free buff has been a pretty solid platform if you want to try out different models for free. The only problem is that it is a little bit slow. But you can compromise that as you get literally hours of free usagefrozen17d ago1.1K201000.27%
Just tested Sonnet 5.5 and It is too good. I told it to create a riding Penguin and the output was great with smooth control, nice graphics. Even the background is so detailed!!! https://t.co/WuqBwC2fVVfrozen23d ago1.1K521031.01%
What the F*** ?? I tried creating the riding Pelican using Kimi K3 max. The funny thing is- it didn't create the Pelican, it just drew the functions. https://t.co/oRYr46bDAlfrozen22d ago1.1K100000.09%
New free model seems to on OpenCode. Gonna drop insights soonfrozen14d ago1.1K500000.47%
Meta cooked hard this timefrozen10d ago1.1K711000.85%
Space Bunny was MiniMax M3.1 Flash all along. The stealth model that topped global leaderboards with 200T+ tokens now has a name and the official M3.1 Flash release is even more powerful, built for agents. → Coding: DeepSWE 74% → Cowork: Automation-Bench 62%, ALE 30% → 1/3 the cost of M3, 3x the speed → Open weights on Oct 23 What stands out is its potential for agentic workflows. With strong coding and automation benchmarks, M3.1 Flash looks well-suited for agents that need to execute multi-step tasks, not just generate answers. And at 1/3 the cost and 3x the speed of M3, running these workflows becomes much more practical. Frontier-level intelligence at bunny prices. 🐰3d ago883401000.57%