Tweet épinglé
BLACKBOX AI is now a @Microsoft Official Partner.
Enterprises can purchase @blackboxai licenses directly through the Microsoft Marketplace.
Built for teams that ship fast and stay secure.
English
BLACKBOX AI
6.1K posts
BLACKBOX AI
@blackboxai
Building Full Self Coding.
San Francisco Inscrit le Ekim 2021
172 Abonnements21.9K Abonnés
BLACKBOX AI retweeté
BLACKBOX AI saw this pattern in February.
Their Claudex Mode: Claude implements, Codex verifies and tests. Same workflow, two models, different roles.
They didn't wait for official permission to compose these systems.
They read the architectural momentum and built toward it.
English
New release:
- Free claude code and codex inside Blackbox VSCode with unlimited agent requests with Minimax
English
@VarunR25117 @files_777 no need for the blackbox api key. Once you have clicked on Claude Code > Minimax, then you should be good to start sending a task to your agent
English
You should probably read this again - it's claude code or codex (coding agents) to be used with the minimax model (which is free)....
with a Pro plan you can use premium models obviously, AND you can even use claude code with openai models and codex with opus 4.6, which is pretty cool
English
@files_777 VSCode Extension, CLI and IDE - Make sure to restart the extension on VSCode just in case 😎
English
@blackboxai Where is this offer available: “Free Claude Code and Codex inside Blackbox VS Code with unlimited agent requests using Minimax”?
Is it available in the Blackbox CLI, VS Code extension, or Blackbox IDE only?
English
@rezoundous @DTK1281422 we have literally as a built in workflow
/workflow lets you set up sessions where claude code and codex cli collaborate to achieve the best task execution
English
100% no need to pick only one
we support claude code and codex cli to collaborate on each task or to compete and have a judge layer pick the best execution
try the Blackbox CLI here blackbox.ai/cli
here is a brief demo x.com/blackboxai/sta…
Tyler@rezoundous
After using Codex and Claude separately for a few weeks, I find that it's not Codex vs Claude. It's Codex AND Claude.
English
@rezoundous 100% no need to pick only one
we support claude code and codex cli to collaborate on each task or to compete and have a judge layer pick the best execution
try the Blackbox CLI here blackbox.ai/cli
here is a brief demo x.com/blackboxai/sta…
BLACKBOX AI@blackboxai
Blackbox multi agent (codex, claude code, blackbox, goose etc...) working together the same task vs. claude code alone
English
@imsroch @steipete no need to pick only one
we support claude code and codex cli to collaborate on each task or to compete and have a judge layer pick the best execution
try the Blackbox CLI here blackbox.ai/cli
here is a brief demo x.com/blackboxai/sta…
BLACKBOX AI@blackboxai
Blackbox multi agent (codex, claude code, blackbox, goose etc...) working together the same task vs. claude code alone
English
claude,
@steipete change my mind to switch to codex
Tyler@rezoundous
Are you team Claude or Codex?
English
New release on VSCode in the next few hours:
- Run Claude Code with OpenAI models
- Run Codex with Anthropic Models
English
Blackbox multi agent (codex, claude code, blackbox, goose etc...) working together the same task vs. claude code alone
English
Tag any agent any model on any task and send it!
/claude
/codex
/gemini
/{agent}
English
New release:
Build and run tests before raising a PR or merging any PR
English
@0xlelouch_ that's because you optimise for the common case, not the best case.
English
If gRPC is faster and more efficient than REST, why not convert all our APIs to gRPC?
English
@anuraggoel you can kill PRs, but you can’t kill validation
we just move it upstream with our multi agent evaluation
x.com/blackboxai/sta…
BLACKBOX AI@blackboxai
New release: Multi-agent part 1
English
AI is quietly deprecating GitHub.
Agents do not need branches, PRs, or CI/CD rituals. They want to ship code straight to the cloud.
The rsync renaissance is here.
High availability. Zero bloat. Faster loops.
English
@bindureddy why picking agents when you can just run all of them and let them fight it out .
English
There are literally a million LLMs but only these eight are useful in the real world
Opus 4.6 - king of coding
GPT 5.4 - everyday use
Grok 4.2 - explains things on X
Gemini Flash - cheap, all purpose
5.4 Thinking - xls and deep research
GLM - cheap coding agents
Kimi K2.5 - claw like agents
MiniMax 2.7 - more claw
Four expensive and four cheap ones.
English
@kepano well yeah, most of these fail because there’s no system to sustain them
English
