BottleCapAI

Making efficiency-first LLMs & Applications.

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Gemma-4-26B-A4B is out!
The most token-efficient Gemma-4-26B-A4B available.

🚀Up to 69% fewer thinking tokens, 38.9% on average
🎯Negligible accuracy loss 0.15pp
🏢ThinkingCap PRO: up to 87.6% fewer thinking tokens, 62% on average - Contact our enterprise[at]bottlecapai[dot]com

📄Blogpost: https://t.co/pFbIEOlwOl
🤗Hugging Face: https://t.co/KTuMSr3UIe
🔗GGUF: https://t.co/D3uRUn3gsf

[BottleCapAI @BottleCapAI]

@garrytan We'd like to enter the chat 👀

[Lupa.cz @Lupacz]

Mikolov a Beck vylepšili umělou inteligenci Googlu. Je zdarma a výrazně méně náročná
https://t.co/bmU1WuOuqq

[Seb Johnson @SebJohnsonUK]

One of the most interesting AI labs in Europe is being built out of Prague

The team believe that frontier models drastically overthink, and therefore burn tokens unnecessarily

So @BottleCapAI (named after Europe's greatest innovation) takes an existing open model and trains it to think less while staying just as accurate.

Last week the lab released 3 models:

>> Nvidia’s Nemotron: up to 31% and on average 23% less thinking, with comparable benchmark accuracy

>> Alibaba’s Qwen3.8-Flash-Next: spending 39% fewer thinking tokens on average for about a point of accuracy loss

>> Google’s Gemma 4: their first ThinkingCap on a Google model, with 39% fewer thinking tokens on average for just 0.15pp of accuracy

The impact of what they're building is less token spend and less hardware is required.

VERY COOL

@JaroslavBeck @itsdavidherel

[BottleCapAI @BottleCapAI]

@SebJohnsonUK 🚀

[BottleCapAI @BottleCapAI]

PRO versions also available: enterprise[at]https://t.co/VoGxKXR5uP

[BottleCapAI @BottleCapAI]

The ThinkingCap series keeps growing.
Last week we released:
→ Nemotron 3.5 Lightning: up to 2.3x more requests served at once
→ Qwen3.8-Flash-Next: up to 51% fewer thinking tokens
→ Gemma 4 26B: up to 69% fewer thinking tokens

Have fun!
https://t.co/t5g7z4hIhn

[BottleCapAI @BottleCapAI]

@SaraJChipps @ChaosAgentsAI 🙌🙌🙌

[BottleCapAI @BottleCapAI]

@codorniou Officially, we can call it a release hattrick! 🧢🚀
https://t.co/yIXn55ivot

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Gemma-4-26B-A4B is out!
The most token-efficient Gemma-4-26B-A4B available.

🚀Up to 69% fewer thinking tokens, 38.9% on average
🎯Negligible accuracy loss 0.15pp
🏢ThinkingCap PRO: up to 87.6% fewer thinking tokens, 62% on average - Contact our enterprise[at]bottlecapai[dot]com

📄Blogpost: https://t.co/pFbIEOlwOl
🤗Hugging Face: https://t.co/KTuMSr3UIe
🔗GGUF: https://t.co/D3uRUn3gsf

[BottleCapAI @BottleCapAI]

Release hat-trick 🚀🚀🚀 3 releases in 3 days!

ThinkingCap isn't just for Qwen. It works across model families, from Alibaba's Qwen to NVIDIA and Google:

Day 1 → NVIDIA Nemotron-3.5-Lightning-30B-A3B
https://t.co/kxy1p0t11O

Day 2 → Qwen3.8-Flash-Next
https://t.co/f3cdQlxnRL

Day 3 → Google Gemma-4-26B-A4B (you're here)

Need even fewer thinking tokens? ThinkingCap PRO goes further, tuned to your workload. 📧 enterprise[at]bottlecapai[dot]com

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Qwen3.8-Flash-Next is out!
The most token-efficient Qwen3.8-Flash-Next available.

🚀Up to 51% fewer thinking tokens, 38.5% on average
🎯Negligible accuracy loss 1.03pp
🏢ThinkingCap PRO: 60% fewer thinking tokens on average - Contact our enterprise[at]bottlecapai [dot]com

📄Blogpost: https://t.co/6xCnG7QyJv
🤗Hugging Face: https://t.co/vbnwfCcv4v
🔗GGUF: https://t.co/ZtQftZpFFy

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Gemma-4-26B-A4B is out!
The most token-efficient Gemma-4-26B-A4B available.

🚀Up to 69% fewer thinking tokens, 38.9% on average
🎯Negligible accuracy loss 0.15pp
🏢ThinkingCap PRO: up to 87.6% fewer thinking tokens, 62% on average - Contact our enterprise[at]bottlecapai[dot]com

📄Blogpost: https://t.co/pFbIEOlwOl
🤗Hugging Face: https://t.co/KTuMSr3UIe
🔗GGUF: https://t.co/D3uRUn3gsf

[BottleCapAI @BottleCapAI]

📷 Same prompt. Same hardware. Who gets there first?
ThinkingCap: Gemma-4-26B-A4B vs. the original, head to head on:
🌟Astrophysics: multiple-choice science
🐍Python: coding
📐Geometry: competition maths
🧬Biology: multiple-choice science

[BottleCapAI @BottleCapAI]

@kusurichan_ 🙌

[BottleCapAI @BottleCapAI]

@almosttomorrowe In our blog post, we are showing how does the accuracy against the same thinking token budget looks like for the original and the ThinkingCap model. And in the image it can be seen that the ThinkingCap model has higher accuracy for the same budgets.

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Qwen3.8-Flash-Next is out!
The most token-efficient Qwen3.8-Flash-Next available.

🚀Up to 51% fewer thinking tokens, 38.5% on average
🎯Negligible accuracy loss 1.03pp
🏢ThinkingCap PRO: 60% fewer thinking tokens on average - Contact our enterprise[at]bottlecapai [dot]com

📄Blogpost: https://t.co/6xCnG7QyJv
🤗Hugging Face: https://t.co/vbnwfCcv4v
🔗GGUF: https://t.co/ZtQftZpFFy

[BottleCapAI @BottleCapAI]

@codorniou 🏎️

[BottleCapAI @BottleCapAI]

@sethforprivacy Let us know! 🙌

[BottleCapAI @BottleCapAI]

🚨ThinkingCap: Qwen3.8-Flash-Next is out!
The most token-efficient Qwen3.8-Flash-Next available.

🚀Up to 51% fewer thinking tokens, 38.5% on average
🎯Negligible accuracy loss 1.03pp
🏢ThinkingCap PRO: 60% fewer thinking tokens on average - Contact our enterprise[at]bottlecapai [dot]com

📄Blogpost: https://t.co/6xCnG7QyJv
🤗Hugging Face: https://t.co/vbnwfCcv4v
🔗GGUF: https://t.co/ZtQftZpFFy

[BottleCapAI @BottleCapAI]

@JaroslavBeck 🚀

[BottleCapAI @BottleCapAI]

💪Side-by-side comparison
ThinkingCap: Qwen3.8-Flash-Next vs. Original model