Codebuff ретвитнул
Codebuff
323 posts

Codebuff
@CodebuffAI
Better agents. Better code. npm i -g codebuff
Your terminal Присоединился Ekim 2024
16 Подписки2.1K Подписчики
Codebuff ретвитнул

We upgrade our file picker agent to use Gemini 3.1 Flash Lite!
It's 4x faster than Grok 4.1 Fast was.
Google DeepMind@GoogleDeepMind
Gemini 3.1 Flash-Lite has landed. It’s our most cost-efficient Gemini 3 series model yet, built for intelligence at scale. Here’s what’s new 🧵
English
Codebuff ретвитнул

Codebuff ретвитнул

NEW Claude Code & OpenCode KILLER! This Just Fixed 90% of AI Coding! (Open Source): youtu.be/4H_an5xtyr0

YouTube
English
Codebuff ретвитнул

Recently saw a video by @piyushgarg_dev on @CodebuffAI and I wanted to try it out but I didnt want to make anything generic so I asked it to make a Temple usin three JS and the outcome is crazy!
English
Codebuff ретвитнул

I found a better terminal agent that beats Claude on benchmarks.
Personally, I tried fixing a complex bug: it got it done in 4 minutes, whereas Claude took around 8 minutes for the same task.
@CodebuffAI

English
Codebuff ретвитнул
Codebuff ретвитнул

Just tried Codebuff, a really great alternative to Claude Code, works a lot faster ⚡️ and burns less credits!
You can try it too + get 500 free credits on using my referral link:
codebuff.com/referrals/ref-…

English
Codebuff ретвитнул

We just switched @CodebuffAI free mode to use MiniMax M2.5.
On our benchmark, the score increased substantially:
54% => 75%
Major upgrade!
Previously we were using GLM 4.7 + Grok 4.1.
MiniMax (official)@MiniMax_AI
Introducing M2.5, an open-source frontier model designed for real-world productivity. - SOTA performance at coding (SWE-Bench Verified 80.2%), search (BrowseComp 76.3%), agentic tool-calling (BFCL 76.8%) & office work. - Optimized for efficient execution, 37% faster at complex tasks. - At $1 per hour with 100 tps, infinite scaling of long-horizon agents now economically possible MiniMax Agent: agent.minimax.io API: platform.minimax.io CodingPlan: platform.minimax.io/subscribe/codi…
English
Codebuff ретвитнул
Codebuff ретвитнул

Just ran Opus 4.6 vs 4.5 on our coding agent evals.
It's noticeably better and costs about 10% more, likely due to more token usage.
Opus 4.6
Score: 8.54/10
Cost: 1.578
Duration: 476 sec
Opus 4.5
Score: 8.27/10
Cost: 1.4118
Duration: 445 sec
Opus 4.6 is live in @CodebuffAI!
English
Codebuff ретвитнул

"I Deleted Claude Code After Finding This!"
Great deep dive on why @CodebuffAI's batteries-included approach is better.
youtube.com/watch?v=y_6hlg…

YouTube
English
Codebuff ретвитнул

Since the last few days, my MacBook Pro battery has been going from 100% to basically 0% in 2-3 hours running Claude Code.
This is when I came across @CodebuffAI, which provides a similar, if not better, experience at using Opus 4.5 for my local coding tasks.
Video out now

English
Codebuff ретвитнул

@CodebuffAI on max mode and best of "n" is addictive to watch! Best CLI UI I have seen!
English

Our old tagline was:
"Make your terminal write code for you."
Let's just say that since we wrote that, having your terminal write code for you has become a lot more popular.
Codebuff@CodebuffAI
New tagline
English







