← Back to leaderboard
bhavani_00007

Bhavy☄️

Verifiedactive
@bhavani_00007

AI | SaaS | Startups | DM for collabs 📩

RosterAI Model News & ReleasesHumor & Meme AccountsTech Career & Coding Education
7,659 followers · 1,860 following · Last polled 11h ago
Perf. Score
66.5
Rank #93
Value Score
89.9
#25 of 273 priced
Median views
13K
p25 6.1K · trailing 7d
Followers
7.7K
+1.0% 7d
ER (impr.)
1.52%
eng ÷ impressions
ER (foll.)
5.72%
eng ÷ followers
Est. CPM
$2
QT-rate basis
WoW views
+23.0%
▲ rising

Rates & value · est. from 13K median organic views

underpriced
Quote tweetbasis
$25
$2 / 1K views
Post
—
no CPM
Thread
—
no CPM
Retweet
—
no CPM
Value Score 89.9 · rank #25 of 273 priced$152.44 / 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 (13K), not the mean (29K) — so a single viral post doesn't inflate a creator's reach. The p25 floor (6.1K) is the level 75% of posts clear.

Recent posts · ranked by views

PostPostedViewsLikesRepostsRepliesQuotesBookmarksER
I asked Opus 5.5, GPT-6 Astra, GPT-6 Sol, Kimi K3 to build the Eiffel Tower in Three.js. Opus 5.5: $8.95 | 10 mins GPT-6 Astra: $7.45 | 8 mins GPT-6 Sol: $3.90 | 5 mins Kimi K3: $4.04 | 6 mins Opus 5.5 really impressed me. The frontend and 3D detailing are seriously good, and I preferred it over Astra for this task. GPT-6 Sol was a bit disappointing, while Kimi K3 looks really good for the price. It delivered a strong result at almost half the cost. Kimi K3.1 is going to be a monster for sure. Overall, Anthropic made a great comeback with Opus 5.5.frozen29d ago167K74335114161800.65%
4 Argon vs Opus 5.5 vs GPT 6 Astra vs Kimi K3 I asked all four to build a 3D whale breaching scene. Argon's level of detail is on another level, and it wins on photorealism. Opus and Kimi look more realistic than Astra. But what a comeback from Google!! 🔥 https://t.co/adz2OExpg1frozen9d ago130K
718
66
53
16
162
0.78%
DeepSeek V4.1 Flash 1017 vs GPT 6.1 Sol I tested this rocket launch scene and DeepSeek absolutely cooked with this model. DeepSeek : $0.78, 3 mins GPT 6.1 Sol: $6.10, 5 mins Nearly 8x cheaper than Sol, and faster too. This thing is a beast. https://t.co/XuODARgwBCfrozen5d ago108K1.5K7562212351.78%
Mistral Large 4 vs Kimi K3 Mistral Large 4 is disappointing. It’s nowhere near Kimi K3. Honestly, it’s way too slow. After an hour of testing, this is what I got. I think even Luna could generate something better than this. But anything is better than nothing. I’d really love to see more European AI models in the AI race.frozen16d ago62K349215411700.82%
MiniMax M3.1 Flash is actually insane. I asked it to build this whale breach scene using Claude Code as the harness, and it cost just $0.25. That's roughly 8x cheaper than Astra and Sol. I didn't expect MiniMax to be this good at 3D. In my testing, it even beats GPT 6.1 Sol, and the result looks better than the others while costing less than DeepSeek Flash’s version. What a comeback by MiniMax.3d ago50K3798432270.93%
I’m done with Muse Wonder. I tried really hard to get good results and kept expecting it to deliver something better, at least once. But that’s just not happening. Sorry, Meta. Based on my tests, it performs below Opus 4.8 but costs as much as Opus 5.5, to be precise. It’s nowhere near Kimi K3, GPT-6.1 Sol, or Astra.frozen10d ago47K296135811400.89%
WTH is Muse Wonder trying to do? I asked it to build a whale breaching scene, and it gave me this lol. I honestly expected a lot more. It’s pretty token hungry too. This cost me $3.20 Kimi, 3 month old model, built something better for just $2.50. Maybe 3D and frontend development aren’t its thing I’ll run a few more tests before judging it.10d ago44K824286160.31%
GPT-6.1 Sol isn’t that good I asked it to build the Eiffel Tower in Three.js, and this is what it came up with. It performs pretty well on some other tests, but it failed badly on this scene compared to other models. For me, it still feels below Astra and even Kimi in frontend and 3D. OpenAI says it’s more efficient, but it seems to consume a lot of tokens.frozen23d ago39K1493291280.54%
Muse Wonder is disappointing. I asked it to build this rocket scene, and it burned $14.3 for this? WTH The result looks average, not even better than Kimi, which built it for around $6.95. Muse Wonder is insanely token hungry. I don't know if this is a temporary bug, but if this is normal, that's concerning. also I'm trying to test Muse Wonder Full now, and I'm wondering how quickly it'll burn through my usage Meta, please fix this.frozen10d ago39K22711449340.84%
4 Argon vs Opus 5.5 vs GPT 6 Astra vs Kimi K3 I asked Argon to generate this rocket scene. Argon focuses more on detail, and I'm loving it. The result isn't as good as Opus or Kimi, but the realism is still there. and Its efficient. I never expected this from Google. https://t.co/ometlpgRq1frozen9d ago25K32526282741.80%
Haiku 5.5 cooked GPT 6 Luna. I asked both to generate a rocket launch scene, with both at max level. Haiku 5.5: $1.30 | ~2 mins GPT 6 Luna: $1.50 | ~2 mins Haiku 5.5 is a huge step up from 4.5. I honestly didn’t expect Haiku to generate a scene this realistically. I’ll attached the Haiku 4.5 result in the replies.frozen15d ago18K1512351191.16%
MiniMax M3.1 Flash vs DeepSeek V4.1 Flash 1017 I asked both to build a 3D rocket launch in Three.js. MiniMax: $0.55 DeepSeek: $0.78 MiniMax came out cheaper, with a cleaner liftoff and a bigger flame trail. It understands intent really well, and it’s fast. I’d definitely love to use it for coding. Cheaper than DeepSeek and better quality. Loving it.2d ago18K2195262171.53%
Gemini 4 Argon vs Opus 5.5 vs Astra vs Kimi K3 I asked Argon to build the Eiffel Tower in Paris in Three.js. Argon’s tower structure is not good. Astra and Kimi come closest to the real Eiffel Tower. Which one looks best to you? https://t.co/85uxyRBzDZfrozen8d ago17K18513432371.62%
GPT 6.1 Sol vs Sonnet 5.5 vs Kimi K3 vs Grok 4.7 I asked all four to build a 3D rose, with each capped at $2. GPT 6.1 Sol took 21 mins. The result wasn’t very realistic, but the colors looked good. Sonnet 5.5 took ~7 mins and looked realistic. Kimi K3 also took ~7 mins and gave the best result IMO. Still setting a high bar for frontend work. I don’t know what’s stopping them from releasing Kimi K3.1. Grok 4.7 took ~25 mins in Grok CLI and gave an okayish result.frozen21d ago16K1224270181.09%
Mistral Large 4 is disappointing. For a model from Europe, this is still reasonably good, and it could be a comeback. European AI has been quiet for ages. It’s near GLM-5.3 but still behind Kimi K3. Still, a big step up from the last Mistral. It’s also pretty expensive compared to Chinese open models, at $1.30 in and $3.90 out. Some of the benchmark scores feel a little benchmaxxed. I’ll test it soon.frozen16d ago14K44025230.51%
Anthropic cooked OpenAI in just one week 😭 I think, GPT-6 Sol is nowhere near Sonnet 5.5 in realism for this scene. Sonnet 5.5 is a huge jump from Sonnet 5, especially in frontend, 3D, speed, and prompt understanding. Honestly, it feels close to Opus 5. https://t.co/qIv3kaEGSCfrozen24d ago13K922150160.94%
DeepSeek V4.1 Flash 1017 is a big upgrade over DeepSeek V4.1 Flash, and it’s super fast. I gave it the whale breach test and compared it with Kimi K3, since it beats Kimi in benchmarks. Its one-shot attempt came out with the colors completely blown out, and the whale wasn’t good (image in replies). So I capped it to $2.50, the same amount Kimi K3 spent on its one-shot, and Kimi still did better. It’s better than V4.1 and GPT 5.6 Sol, but not on par with Opus 5 or Kimi in frontend and 3D.frozen5d ago13K1195261131.31%
Gemini 4 Argon is a beast. I didn’t expect this much realistic detail from any model. The light and shadows, the reflections on the floor, the books on every shelf, the mountains through the glass. All built from pure math in a single browser file. No 3D assets, no game engine. This is tough to beat.frozen9d ago11K30722213884.09%
Fable 5.5 might be very close. A new Claude model just showed up in Artificial Analysis's website code: "claude-onigiri-eap-2" It has four effort levels (low, medium, high and xhigh), and it's set up exactly like Fable 5, Fable 5.1 and Opus 5.5. All of them use the same Opus 4.8 fallback. "EAP" means early access, so Artificial Analysis is probably benchmarking it before launch. Honestly, I think this is Fable 5.5. Same setup as Fable 5.1, same effort levels, and the timing fits. Anthropic usually lets benchmark sites test top models right before release.frozen7d ago11K2035250112.30%
Space Bunny is MiniMax-M3.1-Flash Damn, when I tested it, it was insanely fast, and the results were better than Opus 5 and GPT 6 Sol. We might be getting a model that's fast, efficient, and delivers Opus 5-level performance. 🔥3d ago11K143210071.53%
Space Bunny is amazing. I built this rocket launch scene with it at Max level and compared it with Opus 5, Sol 6, and Gemini 3.8 Flash. It took only around 3 minutes to build this in one shot. Just look at the result. It’s better than Opus 5 at that speed. It clearly mogs Sol 6, Grok 4.7, and Gemini 3.8 Flash.frozen26d ago11K1005251131.36%
OpenAI testing a slow mode with GPT 6.1 Sol? I tested the rocket scene with GPT 6.1 Sol. Yeah, it’s definitely cost-efficient. But it’s way too slow. We got fast mode and now we have slow mode too lol. It took around 30 minutes for me. Not sure if that’s because of the load or if it’s just nerfed. Also, GPT models still struggle with realism. They tend to add same weird color theme. They really need to fix this first.frozen19d ago9.2K983210111.45%
Opus 5.5 vs GPT 6.1 Sol I asked both to build a whale breaching at sunrise in Three.js, each capped at $2. In my opinion, Opus 5.5 won this one, and I’m loving it. Same budget, but Opus finished in ~7 mins and Sol took ~10 mins. Sol’s leap is bigger and splashier, but Opus feels more cinematic to me.frozen12d ago8.8K71118071.10%
Step 5 Preview is now live on OpenRouter I gave 4 models the same 3 tasks: build an Eiffel Tower, a rocket launch, and a building collapse all in Three.js Step 5 Preview came out cheapest and fastest across the board and got good results not expecting this from an open-weight model at all. DeepSeek V4 close behind. GLM 5.3 and Qwen 3.8 took longer and cost more, but still landed solid results. Video says more than I can.frozen14d ago6.1K1068191112.39%
Gemini 4 Argon is releasing today, finally. It is already inside the Antigravity 2.23.0 build Google's own download ships it in three effort levels: Low, Medium and High (gemini-4-argon-low / -medium / -high). https://t.co/PCpvVuJRTffrozen10d ago5.9K1464170102.99%
Artificial Analysis gives Claude models food codenames before launch. Their own site code shows it: • "claude-beignet-eap" = Sonnet 5.5 • "claude-verrine-eap" = Haiku 5.5 • "claude-onigiri-eap" = ? (my guess: Fable 5.5) • "claude-gabagool-eap" = ? (my guess: Opus 6 or a newer Fable 5.5 version) Both new ones were set up like Fable and Opus, then removed from the site today. At least one of them is Fable 5.5.frozen7d ago5.0K72114041.81%
I asked Step 5 Preview to build a survival chase game from scratch. It wrote the whole thing in Three.js, including all the assets. It only cost me $6 and took around 19 minutes. It’s super fast and efficient at the same time. I honestly didn’t expect this level of quality from an open-weight model. Game development is getting really interesting. @StepFun_ai Step 5 Preview is now live on OpenRouter and available in Cline, Kilo Code, Hermes Agent, and OpenCode. Try it: https://t.co/prbNyI4fVHfrozen14d ago4.7K591202132.02%
If this gets integrated into Zoom or Teams, I genuinely don’t know how you’d tell it’s AI 😭 Griffin can see, hear, react and respond in real time, and it can even interrupt you or notice someone walking into the frame. This is getting wild.frozen21d ago4.4K36213081.33%
Holy, Meta is suddenly a real coding contender: Muse Wonder Full beats Claude Opus 5.5 and GPT-6 Astra on long-horizon software engineering, according to Meta's own numbers. Its strongest area is long tasks: - SWE-Marathon: 57.5 vs. Opus 5.5's 46.3 (and GPT-6 Astra's 42.5) - DeepSWE 1.1: 79.0 vs. 74.2 (Opus 5.5) and 74.1 (Astra) So Meta looks best at staying on a long task, not at building from zero. Available now via the Meta Model API and Muse Code. Didn't expect Meta here this soon.frozen10d ago3.3K38314041.78%
Day 1 of Tibo’s 28 day promise is speed Astra and 6.1 Sol should be around 50% faster, going from 30 to 50 tokens/sec. Still not enough to move me off Claude. Hopefully faster doesn’t mean burning through the weekly limit faster I’ll test it and let you guys know in an hour.frozen17d ago2.1K28015002.01%
Gemini-4-Argon is here guysss🥳 https://t.co/PdcnjeWlATfrozen9d ago1.8K31310022.57%
So Space Bunny was MiniMax M3.1 Flash all along I tested it during my 3D model comparisons, and it performed surprisingly well against much more expensive models. That was an older checkpoint, and it still topped OpenRouter for 10 straight days. The new version looks even more promising, especially at just $0.15 input and $0.60 output per million tokens, with speeds of 200+ tokens/sec. MiniMax Code users get unlimited access until Oct 25, and M Plans are 50% off the first month until Oct 31.3d ago1.6K2618002.18%
I built this 3D interactive car store with Space Bunny (a stealth model on OpenRouter) It took just 8 minutes, with only 1 follow-up prompt and $0 API cost. Drive around a tiny island, park at a showroom, pick a paint colour, test drive it, then check out from the driver's seat. This thing is a beast at 3D/frontendfrozen22d ago1.3K35111053.92%
Space Bunny is another stealth model that launched yesterday. I just asked it to generate a modern website and it created this in one shot. WTH, I didn't expect this from a stealth model From my testing, it was fast, finishing in under 4 minutes, feels like it’s at Flash level. It created all the assets, SVGs, and everything directly in code. It also seems to understand the intent behind the prompt really well.frozen27d ago1.3K22111022.75%
I’ve been comparing AI models side by side for months. Same task, very different costs, and the expensive model isn’t always better. Most teams never test this. They pick one big model and let the bill grow. RouterPlus does it for you: the right model for your real usage, a provider that holds up at your scale, and the best price for your workload. Free review for any team spending on AI. https://t.co/ann35hdKSSfrozen9d ago1.1K19011002.73%
everyone’s racing to build smarter models meanwhile half of them can’t even read a webpage properly thank god! https://t.co/VjG0IdIdf7 is fixing this3d ago82028013015.12%
@OfficialLoganK can I get Ultra access? 🥺👉🏻👈🏻 I really wanna test Argon 4 with my prompts9d ago3611402004.43%