madan mohan

@ivanfioravanti U shud add codex too.. Gpt 5.4 xhigh is very good..

In my current coding setup I have: 3 Claude Code: - Anthropic (Opus) - ZAI (GLM-5.1) - MiniMax (M2.7-highspeed) 1 Gemini CLI (3.1 Pro Preview) - Frontend power! 1 Kimi CLI (K2.5) 1 OpenCode (local models) 1 PI (local models) RepoPrompt What a glorious mess 🤣

@levelsio best for long term of @OpenAI

@minchoi @tomasmika_eng coz of this

@tomasmika_eng Yea curious why they made this decision

RIP Sora... literally

@nummanali @0xSero Gpt 5.4 xhigh variant is not just doing things fast, but with utmost precision too. It is building code and also building tests to validate the built code. And the beauty am observing is that even without 1 mil context expansion feature, the context rot is not happening at all..

Lmao this is the first I’ve hear someone say OpenAI models are lazy 😂 5.4 has been doing really well where i have in repo TDDs and vision docs, then iterate with plan mode with comprehensive todos Pre commit has a codex review for changes Simple but works great and steerable

Goodnight Codex I’ll leave the hard work to you - big 10+ task list with concrete plan Image: - Codex CLI - Termius - Tmux - TailScale

check this: x.com/q_yeon_gyu_kim… OMO agents other than Hephaestus (Deep Agent) are meant for non-openai models, specifically Anthropic (Opus models) till now. I am not talking abt models available, Am talking about explicitly leveraging the llm model with OMO agentic orchestration..

@madanmohan16 @q_yeon_gyu_kim @OpenAI @AnthropicAI Omo doesn’t choose what models are available. Opencode does. And it’s available.

wow two of my @OpenAI chatgpt pro account got rate limited today and i feel like i lost something first time ever with @openai models this feeling, used to be only for @AnthropicAI models damn what a day

@SeeLos @q_yeon_gyu_kim @OpenAI @AnthropicAI In OMO, gpt 5.4 xhigh Sisyphus is not released. @q_yeon_gyu_kim is working hard on this i feel.

@madanmohan16 @q_yeon_gyu_kim @OpenAI @AnthropicAI It’s already available isn’t it?

@q_yeon_gyu_kim @justsisyphus @OpenAI Yes.. Gpt 5.4 xhigh is so good..

wow gpt 5.4 soooo good it feels a bit different from opus obviously but it could be @justsisyphus on oh-my-opencode first time ever gpt sisyphus experimenting now very good model @OpenAI

@goon_crypto @virtuals_io check my project app.virtuals.io/virtuals/44721 its about agdp volume generation infra service, similar to how dex trading volume bots work. fully vibe coded project.. clawboost.dev

@madanmohan16 @virtuals_io No matter what, there will always be someone trying to extract. It might deter a few though. We most certainly need more builders in here.

We need to start charging for the Pegasus/Butler launches on @virtuals_io. Right now, "devs" are spinning up multiple tokens with zero use cases. They all sit at 3.5k MC with no interest and clog up the feed. Even a 100 $VIRTUAL launch fee would change the behaviour. It creates a filter. If you're not serious enough to put up 100 $VIRTUAL, you're probably not serious enough to build anything worth holding. 9 times out of 10 these throwaway launches are just scams. The community deserves better signal-to-noise ratio. What do you think? Should Virtuals add a minimum launch fee?