I need to wait for the @VulcanBench if what effort should I use.
OpenAI’s release of GPT-6.2 Astra and GPT-6.2 Sol delivers the #2 model in the Artificial Analysis Intelligence Index and most of the Cost per Task Pareto frontier
@OpenAI's GPT-6.2 Astra is priced at $8/$40, 20% below GPT-6 Astra, and cached input drops to $0.40 from $1.00 per million tokens. GPT-6.2 Sol pricing is unchanged from GPT-6.1 Sol, at $2/$10 per million input/output tokens.
Key takeaways:
➤ Second in Intelligence Index: OpenAI now trails only Anthropic’s Claude Fable 5.2 in the Artificial Analysis Intelligence Index. GPT-6.2 Astra jumps 5 points from GPT-6 Astra, overtaking Gemini 4 Carbon (60) and Claude Opus 5.5 (59) at their top effort settings. GPT-6.2 Sol gains 2 points at 57, placing it on par with GPT-6 Astra at ~15% of the Cost per Task.
➤ Occupies most of cost and output token frontiers: OpenAI models now occupy most of the Cost per Task Pareto frontier, sharing the frontier with only Anthropic and Meta. OpenAI also dominates the Output Token per Task Pareto Frontier, with Claude Fable 5.2 being the only model from a non-OpenAI lab appearing on the frontier.
➤ Strengths in science and knowledge: OpenAI performs strongly in both of our key science-focused evaluations, leading Terminal-Bench-Science 0.1, covering agentic scientific research workflows, and placing second only to Claude Fable 5.2 in CritPt v2 (hard subset), covering physics reasoning. GPT-6.2 Astra performs below only Claude Fable 5.2 in factual knowledge as measured by AA-Omniscience v1.1 Accuracy, and in HLE-Diamond, a reasoning and knowledge benchmark.
➤ Leading in 3D and robotics: GPT-6.2 Astra leads AA-3D at 1318, with Claude Fable 5.2 at 1271 and GPT-6.2 Sol at 1259. This benchmark measures how well models build interactive 3D experiences, and is rated by human graders. GPT-6.2 also leads AA-Robotics v0.1, our research preview measuring how well language models can understand, plan, and act in the real world. GPT-6.2 Astra and GPT-6.2 Sol scored 79 and 78 respectively, ahead of Claude Fable 5.2 at 75. Our team is hard at work adding more difficult tasks, as OpenAI and Anthropic models can now perform the majority of basic tasks in this initial research preview.
➤ No improvement in Coding Agent Index: GPT-6.2 Astra is flat with GPT-6 Astra, and GPT-6.2 Sol loses 3 points relative to GPT-6.1 Sol in the Artificial Analysis Coding Agent Index (Codex harness at max effort). For Sol, the drop is driven by a 14-point drop in SWE-Atlas-QnA-AA, where the model is more likely than its predecessor to omit key facts in its responses. Cost efficiency however does place several effort levels of GPT-6.2 Astra and Sol on the Pareto frontier for Cost per Task.
1
1
26
I've turned the FAA router outage into an explainer for $2.65
1. I just paste in the news hook what happened and my own take as a sysadmin.
2. my AIUGC host(my AI avatar) via @treg_ai
3. Vox style skill scenes using @HyperFrames_
4. Text layer using Opus 5.5 High
5. Cut together with captions.
BREAKING: The FBI has reportedly taken possession of the router the FAA says was behind Friday’s nationwide ground stop.
75
Microsoft 365 Outage Explained with Analogy
Take not this is not me this is my AI avatar and clone's my voice using @FishAudio and using @treg_ai playbook to generate as cheap as possible
I use reverse engineer growth skill for inspiration of my favorite creators such as @AliAbdaal Shorts/tiktok/Reels videos
I can now be a marketing machine because of this workflow.
2
1
187
I had fun making this. So I just pasted the article to @bot and made some additional comments and additional instructions. And I make sure that in order to save tokens, I ask it to plan using Claude Fable 5.2 on my MacBook Pro, and the workers will be Opus 5.5 High. And make sure to use @treg_ai . So, and grill me skill to ask questions what do I want for a content
Here it is
1
1
81
I'm gonna try and if @bot can do this I paste the full article. will report the outcome
for testing I will pick a woman and next will be a man and next will be my face and see the analytics based on my niche.
I use @treg_ai to search viral videos on tiktok,reels, shorts
1
24
Love creating this workflow. I film 1 long video for Youtube then clip it by using @jasonzhou1993 UGC guide using @treg_ai
I use @FishAudio for the voice
I use @t3dotcodes
I use @every Compound Engineering Plugin as suggested by the great @mvanhorn and @morluto REA
Great time to be alive
1
3
170
My first UGC video using @treg_ai thanks to @jasonzhou1993 I just paste this to my agent and follow attached my photo and my own video here it is
On the right, I gave it my face asset which is this and it works its magic using Treg.
By the way, I just use the @AsideAI And so far it's really, really good.
🤖 Made with AI
Trying it now. Kinda feels like @TermiusHQ but its same thing i guess. Big fan of @herdrdev will further test this
okay okay. you want mobile, we get it.
HERE IT IS: herdr for iphone, ipad and android.
see which agents need you, open the real terminal and answer. type, paste a screenshot, or just talk to it.
works with herdr-core. herdr tui isn't supported (yet)
33
okay okay. you want mobile, we get it.
HERE IT IS: herdr for iphone, ipad and android.
see which agents need you, open the real terminal and answer. type, paste a screenshot, or just talk to it.
works with herdr-core. herdr tui isn't supported (yet)
I will always start Low effort on every Model and work my way up if its needed fable 5.2 is almost the same as Opus 5.5 High(which is my default lets see
Replying to @claudeai
In early testing on FrontierCode 2.0, Fable 5.2 at High effort scored about level with Opus 5.5 at Max effort, for roughly half the cost per task. At Xhigh effort, it scored the highest of any released model tested.
Medium is the default in Claude apps and Claude Code.
25
If you are using @NousResearch hermes and want to use your claude sub
hermes plugins install claude-subscription-directsdk
install this. It works
25
Power of @bot I do have a bot called alternate shopper that finds the best deal with same specs. Bought a kettlebell email the store its available and they said it is. Stored by disposable cc in the vault. if the call feature is available in NZ. It will be complete bot end to end interaction.
16
To all my American X friends. Can you tell me what will happen to innovation AI technology data centers etc. based on the result of this election there?
13
Enjoying using @NousResearch Hermes mobile app i might need to wait for 2 days if I keep using today.
12
So cringe to make my first vlog. Not a native english speaker but I dont care I want to share my thoughts on @NousResearch iOs app beta well done @Teknium and the team.
Good thing I did have a chance to try it.
I linked it to my linux machine which has hermes installed and gave my hermes the ability to have a GUI for web browsing
Hermes has an app
Pick any model
Get past blocks
Plugins to install
TestFlight is out
Rooting my TV
Go try it
here is the link: hermes-agent.nousresearch.co…
iOs: testflight.apple.com/join/U8…
Android: appdistribution.firebase.goo…
as of now tesflight for both platforms are full so check the link once in a while.
1
2
92
ok @herdrdev i see you. i've never been excited to use a desktop app and can add machines as long as it has herdr installed i can run @NousResearch hermes on my Linux machine and do REA stuff on the side.
43
using @superwhisper in @dhh Omarchy is really good. good thing I've installed in a Linux Machine
2
4
250





























