Ditto

[Ditto @dittops92]

We have released the SOTA opensource model for code generation, Code Millenials 34B with HumanEval score of 80.48 🚀

The model surpases the humaneval value of Gemini, GPT3.5, Deepseek Instruct and WizardCoder

Model -> https://t.co/bxhxwGzRyI
Github -> https://t.co/uqvxIydI5e

[Ditto @dittops92]

Published a paper on how context compaction improves agent success and accuracy

https://t.co/6ZTHuXWegl

[Ditto @dittops92]

@akshay_pachaar Try this if you like lmstudio type gui for decision models

https://t.co/4AJLJrBPfd

[Ditto @dittops92]

@RishiUvaach Thanks for sharing. Happy to know that you found it interesting.

[Rishi @RishiUvaach]

Found an interesting open-source project: Bud Decision Studio.

Think of it as #LMStudio for decision models. Run #Jev -like System One models locally, experiment with them, compare models, evaluate decisions, and expose them through an API.

Repo : https://t.co/UI1Rwi8R0i

[Bud Ecosystem @BudEcosystem]

#Jev opened up a new category: System One models that decide fast and cost far less.

Today we're open sourcing Bud Decision Studio, the LM Studio for decision models. Run open models like #Laya, #Kev & #CLM on your own hardware.

Mac, Windows, Linux.

https://t.co/MHcKyMxW8E

[Ditto @dittops92]

@claudesfun Looking forward

[Ditto @dittops92]

@OpenRouter @typesafeai For anyone who wants to try System One-style decisions locally: we just open-sourced Bud Decision Studio. 11 open models, GPU or CPU, with an API compatible with Jev/OpenRouter/Vercel AI Gateway. https://t.co/ojCGx2c8bv

[Ditto @dittops92]

@sirshibaninja Agreed, and the open side is catching up. We just open-sourced Bud Decision Studio: run 11 open decision models locally (GPU or CPU) behind a Jev/OpenRouter-compatible API. https://t.co/ojCGx2c8bv

[Ditto @dittops92]

@andrey_gubanov1 Nice numbers. On the native-desktop side, we just open-sourced Bud Decision Studio: 11 open decision models, GPU or CPU, with a Jev/OpenRouter-compatible API. Curious how it compares to your browser runs. https://t.co/ojCGx2c8bv

[Ditto @dittops92]

We just open-sourced Bud Decision Studio, the "LM Studio" for decision models.

It runs locally on GPU or CPU, with 11 open models and an API compatible with TypeSafe, OpenRouter and Vercel.

https://t.co/ojCGx2c8bv

[Bud Ecosystem @BudEcosystem]

Bud AI Foundry just got better at managing AI costs.

The new Cost Optimiser gives you visibility and control over AI spend with intelligent caching, configurable optimisation, cost tracking, and request-level insights.

Watch the walkthrough ↓

#BudAI #EnterpriseAI #GenAI

[Ditto @dittops92]

@steipete @WesRoth That’s not true. Harness play a big part in any agentic usecases, even in security. With model alone it’s not useful.

[Bud Ecosystem @BudEcosystem]

Every boundary in your AI stack is a tax. Four line items, latency, accuracy, tokens, audit, levied at every handoff before any AI runs. You don't have to pay it: the Bud Novaria AI OS ends it at the architecture. https://t.co/qLTkEHaHXD

[Ditto @dittops92]

@Im_IrushiK @AnthropicAI This has been the case for long now

[Bud Ecosystem @BudEcosystem]

𝗪𝗵𝗮𝘁 𝗱𝗼𝗲𝘀 𝗶𝘁 𝗹𝗼𝗼𝗸 𝗹𝗶𝗸𝗲 𝘄𝗵𝗲𝗻 𝗔𝗜 𝗯𝘂𝗶𝗹𝗱𝘀 𝗔𝗜?

Here’s Bud Agent in action.

Give it a requirement in plain language, and watch it go from intent to a production-ready AI agent—handling the models, tools, guardrails, and evaluations along the way.

[Ditto @dittops92]

@bcherny Can’t use fable for everything, it’s costly. Even two 200$ account is not enough to use fable.

[Ditto @dittops92]

@bcherny @kimmonismus I think quality issue of Opus is real. I started facing this after the realesse of fable.

[ClaudeDevs @ClaudeDevs]

MCP 2026-07-28 is live and it's the largest update to the protocol since launch.

MCP is now stateless, making it easier to deploy and scale remote servers.

https://t.co/K8KqxbUh4e

[Claude @claudeai]

Introducing Claude Opus 5.

It's a thoughtful and proactive model that comes close to the frontier intelligence of Fable 5 at half the price.

[Bud Ecosystem @BudEcosystem]

AI was supposed to transform businesses. For most, it has only transformed the size of the subscriptions and bills. Somewhere between the demos and the deployment, the magic disappears.

Why?

Find out below,
and we'd love to hear your thoughts.

#Bud #AIOS