Ten posts selected from the week’s strongest AI news, credible reports, leaks and model progress—shown with their original X media and engagement snapshot.
10 stories839K combined views8.6K likes
Rank#1
LeakNews
GPT-6 will drop soon
Sam Altman is heading to Washington this week to preview OpenAI’s most powerful model yet👀:
• It's capable of original scientific discovery
• It solved an 80 year old maths problem autonomously
• Runs for much longer without constant supervision
• Powerful cyber capabilities (it's likely the model involved in the Hugging Face incident)
• This will have a new focus on “knowledge per dollar” rather than benchmarks alone
This model looks much bigger and better than a normal ChatGPT incremental update.
Are you ready for GPT-6?
Alibaba just launched Qwen-Image-3.0 and it looks ridiculously capable
The official improvements include:
• Up to 4.5K-token prompts
• Legible text as small as 10px
• Native rendering across 12 languages
• More than 100 supported visual styles
• Complex newspapers, exam papers and interfaces generated in one pass
• Entire 3×3 grids containing nine detailed infographics
There are no independent benchmark results yet, but the examples Qwen have shown are incredible, I can't tell if some of the images are AI or real, they may have just built one of the most useful image models available.
The Chinese labs are moving unbelievably fast.
Do you think this beats the competition in image gen?
Composer 3 is reportedly being tested inside Cursor under the codename “Vega”
The leaked checkpoint appears to include:
• Fast mode
• Medium reasoning
• High reasoning
• XHigh reasoning
• Six internal model variants
• A public release possibly coming very soon
Its coding and agent performance should sit somewhere between GPT-5.4 and GPT-5.5, and remember that composer models are ridiculously cheap and incredibly fast.
With how good Composer 2.5 was, are you looking forward to Composer 3?
• Reportedly keeping the same pricing as Fable 5
• Expected to focus heavily on long horizon reasoning and agent work
Fable 5.1 will drop close to the GPT-6 launch so Anthropic will probably try to launch first.
August is looking ridiculous for AI.
Elon Musk says xAI’s next model:
• Uses ~ 2 trillion total parameters
• Is better than the current 1.5T model “in every way”
• Could challenge Kimi K3 and Claude Opus 4.8 👀
• Is expected to finish initial training this week
• Is targeting similar speed and token efficiency to Grok 4.5
In my opinion the most interesting par about it is that Grok 4.6 will benefit from real world engineering data and expertise across Tesla, SpaceX, Starlink and Cursor.
The exact training mixture has not been disclosed yet, but this provides something other AI companies cannot.
If it keeps Grok 4.5’s speed/efficiency while reaching the capability of Kimi k3 / more advanced models, it will become the best value AI model imo, the pedigree of the training data + efficiency + speed will make it incredible.
Could Grok 4.6 become the new best cost/performance model, and what improvements would you want in general for the next model?
GPT-5.6 Sol is entering its Cerebras launch window
OpenAI’s fastest frontier model deployment is scheduled to begin this month:
• Up to 750 output tokens per second
• Full GPT-5.6 Sol intelligence
• Initial access limited to selected customers
• Capacity will expand gradually
• Long coding and reasoning outputs could finish in seconds
This could make a frontier model feel almost instant for coding, research and agent tasks.
Are you looking forward to this rolling out and what will you test it on first?
DeepSeek’s legacy models shut down in less than one hour
The final migration deadline for DeepSeek V4 is almost here:
• deepseek-chat will become inaccessible
• deepseek-reasoner will also stop working
• Deadline is 15:59 UTC
• You need to move to V4 Pro or V4 Flash
• Existing API keys and the base URL remain unchanged
Have you moved your projects over to DeepSeek V4 yet?
Moonshot are releasing the full model files by July 27:
• 2.8T total parameters
• Native vision capabilities
• 1M token context window
• Built for coding, reasoning and long-running agents
• Technical details expected alongside the weights
This will become the biggest open weights AI release ever.
Just look at what has dropped so far:
• Claude Opus 5
• Grok 4.5
• DeepSeek V4
• Kimi K3
• Qwen 3.8 preview
• GPT-5.6 Sol
• GPT-5.6 Terra
• GPT-5.6 Luna
• Gemini 3.6 flash
• Gemini 3.5 Flash-Lite
• Gemini 3.5 Flash Cyber
• Muse Spark 1.1
• Might have missed some too ?
And Kimi K3’s full weights are still due before the month ends.
The pacing of releases is getting ridiculously fast.
Do you think August will top this?
