@theo Harness is what makes these models work really well.
English
Saad Hasnain Khan
91 posts
Saad Hasnain Khan
@105Hasnain
Founder • AI Engineer • Full Stack Developer
Katılım Ekim 2016
212 Takip Edilen53 Takipçiler
"Why do you use gpt-5.6 and kimi k3 inside of Claude Code?"
This vid goes in depth on what Claude Code gets right, what Codex gets wrong, and how to copy my setup :)
Blessed by Tibo btw x.com/thsottiaux/sta…
Tibo@thsottiaux
If you aren't yet bold enough to install the Codex app, you can stay in the presence of your orange crab and point it at GPT 5.6 Sol. Takes 5 minutes. Kudos to Theo for explaining one of the ways to get this done. Step 1: Install CLIProxyAPI Step 2: Connect Step 3: Define following alias and enjoy claudex ``` alias claudex='CLAUDE_CODE_SUBAGENT_MODEL=gpt-5.6-sol \ CLAUDE_CODE_ALWAYS_ENABLE_EFFORT=1 \ CLAUDE_CODE_MAX_TOOL_USE_CONCURRENCY=3 \ ENABLE_TOOL_SEARCH=false \ claude --model gpt-5.6-sol' ``` If this gets blocked, I owe you a reset.
English
One thing to remember: Kimi K3 isnt another Sonnet or GPT-5.5 tier model. its Kimi’s frontier model.
So yes its going to use more tokens, especially on exploration heavy tasks. Thats the tradeoff for using a much larger and more capable model.
English
@105Hasnain @maxedapps Have you gotten a chance to test anything long horizon?
English
@ShellDropSshApp @maxedapps Honestly CLI experience isnt that much upto the mark compared to claude or codex but it has pretty much almost all the features like all the cmds, plugins etc
English
@105Hasnain @maxedapps How do you find the CLI compared to Claude, Grok, Hermes, etc.?
English
@thsottiaux Performance I know you are gonna say fast mode but sometime its still not fast
English
Have anyone tried voice conversation in ChatGPT I am pretty impressed
English
Is anyone able to access /agents in Claude? I believe it no longer exists.
English
@maxedapps This is right! Most importantly you should know about your system like what things need to be catered and ask agents to cater those. Context needs to be clear and your repo should have all the necessary .md files
English
Let me repeat: Agents will ALWAYS find more issues when you ask them for it.
You can have them write the code and let different agents with fresh context review that same code and the "improved code" dozens of times => they will not stop finding issues.
It's YOUR job to determine which issues really are issues. And which potential improvements you really want or need.
English
An image of the text expressible in fewer tokens than the text itself!
I had a friend doing a PhD in cybersecurity who told me a month ago, but I didn’t try it. It is actually true!
English
@leodev Yep dont expect luna to perform well like Fable because its their affordable and lightweight model. Its ideal for high-volume, low-stakes tasks. Clearly even in chart you can see its somewhere around 11%
English
@105Hasnain Yea but this is luna, the really cheap one. Not Sol
English
How is GPT 5.6 Luna Max (the really cheap one), performing as good as Fable 5 on DeepSWE?
I haven't tried it but I really feel like something's up with this...
English
@Rasmic The moment I saw this, I took my car keys, sighed with relief, and headed out. 😮💨
English
WE KEEP SHIPPING
WE KEEP SHIPPING
FABLE 5 IS STILL HERE
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English
Just like you can access Codex from Claude using codex CLI, have anyone tried using Claude from Codex using claude CLI ?
English
@kutrinet @dr_cintas Thats what actually happening behind the scene
English
@dr_cintas Or you just ask Fable to use GPT-5.6 via Codex CLI 🤷🏼♀️
English
OpenAI’s GPT-5.6 Sol model can now run inside Claude Code.
There are two ways to do it, and both take minutes to set up, here’s how:
Option 1, the official plugin:
/plugin marketplace add openai/codex-plugin-cc
/plugin install codex@openai-codex
/reload-plugins, then /codex:setup
That unlocks /codex:review, /codex:adversarial-review, and /codex:rescue. Claude writes, GPT critiques, you ship.
Option 2, the proxy. One alias makes Sol your main model:
alias claudex=‘CLAUDE_CODE_SUBAGENT_MODEL=gpt-5.6-sol CLAUDE_CODE_ALWAYS_ENABLE_EFFORT=1 claude –model gpt-5.6-sol’
Bonus: Claude Code lets you set subagent model and effort yourself.
Sol Ultra runs can burn several times the tokens of a base run, so right-sizing delegated work saves real money.
English
Are we getting a reset this time?
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English
Do you think OpenAI and Anthropic keep extending frontier model access because the “limited-time” fear makes people test more, build more, and give them better data for the next model? 👀
English
As soon as Claude extended their Fable access, Codex planned to remove the 5 hour usage limit 🤯
Tibo@thsottiaux
Morning. The last 48 hours of Codex and ChatGPT Work have been intense! Three important updates: - Temporarily removing the 5 hour usage limit restriction for all Plus, Business and Pro plans - Rolling out changes that will make GPT 5.6 Sol more efficient across the board and that will be reflected in less usage being used so that it can take you further. Exact impact to be quantified and shared - We hit 6M active users, and are landing a usage reset in the next hour Go do things
English
@robj3d3 @AnthropicAI What I do is plan all my tasks in advance, you can use Notion or any other tool then use the routine or scheduling feature to run them overnight, when your usage limits are more likely to reset. Basically, automate your AI usage.
English
I'm done with them fucking with us.
Ended up in hospital today from stress.
Stayed up all night pushing my limits too hard, thinking it would be removed.
Health comes first.
Do better @AnthropicAI
Claude@claudeai
We're extending Claude Fable 5 access on all paid plans, as well as keeping Claude Code’s weekly rate limits 50% higher, through July 19.
English