Jose

phyton dev.

[Jose @josedevau]

How we got Ornith-1.5 trending 📈

4.4M impressions, 1.6K retweets, 6.7K likes on a single launch post.

We didn't just announce a model, we told a story, positioning that stood out from every other AI launch that week.

This is what AMA does for tech & AI companies.

[Ornith @ornith_]

Aloha! 🌺Introducing Ornith-1.5, a family of open-source LLMs spanning 9B Dense, 35B MoE, and 397B MoE, trained with self-improving strategies.

It achieves state-of-the-art performance among open-source models of comparable size and delivers performance comparable to Claude Opus 4.8 across reasoning, agentic, and coding tasks:
✅Terminal-Bench 2.1 (86.1)
✅SWE-Bench (86 on verified, 65.1 on pro, 79.6 on Multilingual)
✅DeepSWE (56)
✅HLE (44.6)
✅ClawEval (81.4)
✅Tool Decathlon (71.2)

Ornith-1.5 takes a major step toward training foundation models through end-to-end self-improvement, extending the self-scaffolding strategies introduced in Ornith-1.0 into a more complete self-improvement loop: the model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve.

All models, along with their quantized versions (FP8, GGUF, MLX, and NVFP4), have been released under the MIT License, enabling unrestricted commercial and research use.
📘Tech Blog: https://t.co/OZ63scRWLB
🤗Huggingface: https://t.co/mGJLwhrQOM

[Ashley @ashleymakes_]

I find it astonishing how the Chinese government is trying to hinder our country's progress just to gain an advantage in the AI market.

[Jose @josedevau]

They're trying to avoid the inevitable; the advancement of AI is something that's going to happen no matter how much they oppose it, whether it's the Chinese government or activists.

[AF Post @AFpost]

China is massively funding activists and influencers to oppose Flock cameras.

The same group funding the “No Kings” anti-Trump protests is tied to anti-Flock content and is led by pro-China billionaire Roy Singham, who backs Hasan Piker, the leftist extremist.

Follow: @AFpost

[Jose @josedevau]

You need motion to stand out. You can't afford a designer.

@TopviewAIhq Motion Studio generates it for $3.

Demo looks $20k. You spent $3.

[TopviewAI @TopviewAIhq]

$3 to create motion graphics that look like a $3,000 After Effects project?

That’s Topview Motion Studio. Powered by Seedance 2.5.

No timelines. No keyframes. No motion designer required.

Just upload your references, describe your idea, pick a style, and generate a polished motion video in minutes.

#TopviewAI #AIVideo #AIMotionVideo #AIContentCreation #AIMarketing

[Jose @josedevau]

@Dellafuentee_ https://t.co/UReCvL3OMD

[Jose @josedevau]

How we got Ornith-1.5 trending 📈

4.4M impressions, 1.6K retweets, 6.7K likes on a single launch post.

We didn't just announce a model, we told a story, positioning that stood out from every other AI launch that week.

This is what AMA does for tech & AI companies.

[TopviewAI @TopviewAIhq]

$3 to create motion graphics that look like a $3,000 After Effects project?

That’s Topview Motion Studio. Powered by Seedance 2.5.

No timelines. No keyframes. No motion designer required.

Just upload your references, describe your idea, pick a style, and generate a polished motion video in minutes.

#TopviewAI #AIVideo #AIMotionVideo #AIContentCreation #AIMarketing

[Jose @josedevau]

@AliceInTheData 🔝

[Jose @josedevau]

How we got Ornith-1.5 trending 📈

4.4M impressions, 1.6K retweets, 6.7K likes on a single launch post.

We didn't just announce a model, we told a story, positioning that stood out from every other AI launch that week.

This is what AMA does for tech & AI companies.

[Ornith @ornith_]

Aloha! 🌺Introducing Ornith-1.5, a family of open-source LLMs spanning 9B Dense, 35B MoE, and 397B MoE, trained with self-improving strategies.

It achieves state-of-the-art performance among open-source models of comparable size and delivers performance comparable to Claude Opus 4.8 across reasoning, agentic, and coding tasks:
✅Terminal-Bench 2.1 (86.1)
✅SWE-Bench (86 on verified, 65.1 on pro, 79.6 on Multilingual)
✅DeepSWE (56)
✅HLE (44.6)
✅ClawEval (81.4)
✅Tool Decathlon (71.2)

Ornith-1.5 takes a major step toward training foundation models through end-to-end self-improvement, extending the self-scaffolding strategies introduced in Ornith-1.0 into a more complete self-improvement loop: the model proposes new tasks, generates task-specific scaffolds, and produces solution rollouts for reinforcement learning, continuously creating new learning experiences from which it can improve.

All models, along with their quantized versions (FP8, GGUF, MLX, and NVFP4), have been released under the MIT License, enabling unrestricted commercial and research use.
📘Tech Blog: https://t.co/OZ63scRWLB
🤗Huggingface: https://t.co/mGJLwhrQOM

[Jose @josedevau]

@alxcz_dev 🔝

[Jose @josedevau]

@dereckivxn 🔝

[Jose @josedevau]

@elsalearnai 🔝

[Jose @josedevau]

@AIByLeo 🔝

[ALICE ⚡ | AI Growth Mentor @AliceInTheData]

MIT license across 9B, 35B, and 397B variants, plus FP8/GGUF/MLX/NVFP4 quantized versions, that’s a genuinely open release, not just a benchmark flex 🏆

[Leo ☄️ @AIByLeo]

Ornith-1.5-35B activating only 3B params per token yet beating dense models like Gemma 4-31B is the kind of efficiency gain that actually matters

[Dereck. M @dereckivxn]

Wild seeing an open-source model family hit numbers comparable to Claude Opus 4.8 across reasoning, coding, and agentic tasks. Ornith-1.5 is worth a look.

[Jose @josedevau]

@MikuBTC Very good analysis. Let’s just wait.

[Jose @josedevau]

Anthropic has an 80% chance of IPOing by October 31, 2026, per Polymarket:

[Dereck. M @dereckivxn]

Coding is easy mate

[Dereck. M @dereckivxn]

Vibe coding feels cheap when it’s one person burning tokens.

Put 500 devs on agents and picking the wrong model gets expensive very fast.

[Jose @josedevau]

The circular nature of AI deals, per Bloomberg:

[Jose @josedevau]

8% chance OpenAI's valuation hits $2.5 trillion by December 31, per Polymarket: