The A.I. Daily
Bindu Reddy@bindureddy
The US and European models claim to be GLM 5.2 level…. Just when GLM 5.5 is gonna ship It will take 3-6 months for western open source to catch up to Chinese models 🤣
Bindu Reddy
170 storiesBindu Reddy is an AI commentator who posts on X about model capabilities, pricing and where the industry is heading. Recent posts focus on Google's newly announced Carbon model, including his LiveBench expectations, pricing speculation and a rumored OpenAI model called Bel, plus his purchase of Google shares. He also covers Argon's weak instruction-following, Chinese models distilling Astra, Anthropic's reliance on API revenue, and AI moving from coding to physics.
Bindu Reddy@bindureddy
There is a solid chance Astra and Fable will be much cheaper 🚀 Super excited for this as this will truly bring super intelligence to the masses It’s incredible how under hyped these models are because they are expensive
Bindu Reddy@bindureddy
Chinese open-source models are exceptionally good at easy tasks No US model or frontier lab comes close DeepSeek is still the best option for simple personal agents. The inference per user costs almost nothing!
Bindu Reddy@bindureddy
Next big open source drop - GLM 5.5 Of course they will claim it is Fable 5.1 level 😂 I suspect it will beat Opus and Sol though!!
Bindu Reddy@bindureddy
- Sol 6.1 is excellent at browser use - Opus 5.5 is pretty good at raw coding Combined together, they make the best software
Bindu Reddy@bindureddy
Vibe coders should rejoice The hardcore techies made fun of them and insisted on using IDEs and CLIs Now everyone simply uses a chat box and talks to an agent 🤷♀️
Bindu Reddy@bindureddy
Alarmingly, OpenAI doesn’t seem to be very focused on models They should be - the came back because of Astra and Sol - both solid models It’s time to ship a truly better model
Bindu Reddy@bindureddy
Multiple US open-source models dropping soon Axiois claims that the billions invested in the US stealth foundation model start-ups will pay off Apparently they are launching open-weight models that with some tweaking will be better than frontier models Skepti…
Bindu Reddy@bindureddy
OpenAI is confirming the Astra 6.1 launch for next week A great time to also cut Astra’s price and push for adoption Models are going to self improve rapidly and become a lot more efficient
Bindu Reddy@bindureddy
Open source models have a very short window of 8-12 weeks They will either get vastly better and enter the self-improving loop phase or will be left behind The frontier labs are already in the loop phase Expecting some big open source announcements soon
Bindu Reddy@bindureddy
OpenAI says they are launching again next week It would be absolutely astounding if it were Astra 6.5 I am pretty sure they can nail this! 💪
Bindu Reddy@bindureddy
Gemini 4.0 Pro is a BIG WIN for Google - It's 5x cheaper than Astra - even at 95% of Astra's performance, it's still frontier intelligence It has finally put Google ahead of open-source
Bindu Reddy@bindureddy
PREDICTION - Two Chinese model drops in the next couple of weeks All claiming they are better than Astra, Sol and Opus 5.5
Bindu Reddy@bindureddy
Rumors about SSI, Ilya's start-up are swirling They have a breakthrough to announce Hopefully something all of us can use 😃
Bindu Reddy@bindureddy
Fable 5.5 release is impending - as soon as next week Insanely exciting as this will by far be the most capable model accessible to human kind!! Making extremely innovative and amazing things possible 🔥
Bindu Reddy@bindureddy
Gemini 4 Pro Vibes So Far - Google’s best model - Astra-level on 3D games - expect it to be super fast - lower hallucination than Opus models Seems quite promising….
Bindu Reddy@bindureddy
Moved back 100% of our coding agents to Fable 5.1 It’s heads and shoulders above the rest and requires the least supervision Still, it’s far from perfect! Can’t wait for Fable 5.5
Bindu Reddy@bindureddy
Gemini Argon's pricing is the best part of the announcement It is 5x cheaper than Astra As with all Google models, it looks great on paper. The benchmarks are astounding. 😂
Bindu Reddy@bindureddy
Sol 6.1 day two vibes - solid model - excellent for scheduled tasks - good for browser use - excels at data analysis Usage should sky rocket given the price
Bindu Reddy@bindureddy
CODE RED FOR THE REST OF THE WORLD Super Intelligence can literally make magic; it built a Sol 6.1 in less than a week BAD NEWS - only OpenAI and Anthropic have access to it Expect both of them to be $10T companies in 12 months and a MONSTER duopoly. Other su…
Bindu Reddy@bindureddy
GPT SOL 6.1 is AGI for the masses Absolutely remarkable that OpenAI shipped a new version in less than a week Their big model (Bel) that is training these small models must be astounding 🤯
Bindu Reddy@bindureddy
Indeed! Sol 6.1 Is As Good As Astra On Benchmarks Plus, it is 5x cheaper!! OpenAI is totally back in less than a week
Bindu Reddy@bindureddy
So far, the best thing about Dev day is the new model - 6.1 Sol Would have been way more compelling if it was 6.1 Astra
Bindu Reddy@bindureddy
GPT-6.1 Sol is here - claims to be as good as Astra - $2 / $10 for 1M input / output It will be amazing even if it beats 5.6 Sol 😅
Bindu Reddy@bindureddy
🚨 Self Improving Personal Agents - chat with them on WhatsApp and Telegram - persistent memory - learns how you do work - automates all the heavy lifting
Bindu Reddy@bindureddy
Sonnet 5.5 Is Not A Very Good Model - spins on max mode - underperforms on agentic coding - slightly worse than Terra - stick either Sonnet 4.6 or DeepSeek Flash Anthropic should retire Sonnet and Haiku, release their big Model-2
Bindu Reddy@bindureddy
Confirmed! OpenAI is going to follow Muse, Grok, and 2 dozen start-ups.... They will be releasing a personal agent WAY MORE IMPORTANTLY They will likely pre-announce their big new model, Bel and will time it with Gemini 4.0 Pro Gemini, of course, wants to su…
Bindu Reddy@bindureddy
🚨 Combine Astra 6 and Opus 5.5 for 3D games and apps - Picks the best model based on the prompt - Set up complex multiplayer apps - Create entire 3D platforms with payment - Always on Abacus AI Super Computer
Bindu Reddy@bindureddy
List of things OpenAI can do on dev day - release Astra 6.1 or 6.5 - commit to releasing Bel: bigger than Astra - cut Astra’s price by 50% - cult SOL 5.6 (excellent model) by 50% And yeah a bunch of rando features on Codex but the models matter a lot more
Bindu Reddy@bindureddy
Deepseek is the literal king of acing benchmarks Their latest Flash model claims to be as good a Fable and Astra Anyone who has used it for like two minutes knows it’s not true 🤣