OpenAI GPT-5.6, Google Gemini 3.5, And Dozens Of Open-Source Models Launch Soon

Ah! Everything is timed to perfectly launch together

- GPT 5.6 - Fable 5 - Gemini 3.5

Plus about a dozen open source models! All timed to launch in 15 days 🚀

@bindureddy That's too long in AI years.

@bindureddy Hope they will release for everybody this time.

@bindureddy still won't run on my phone

@bindureddy Sonnet 5?

@bindureddy and none of the normies can use them. perfect.

@bindureddy When?

@bindureddy wishful thinking

@bindureddy prediction based on extrapolating the past:

fable 5 and gpt 5.6 launch within hours of each other, and everybody uses both of them and raves about how great they are.

gemini 3.5 demonstrates great benchmarks but messes up the harnesses and APIs so bad that no one uses it.

@bindureddy I’m honestly most excited about the open source models: DeepSeek 4.1 with real multimodality, GLM 5.5...

@bindureddy the real unlock isn't just more models, it's the infrastructure that lets them be used at scale. someone's going to figure out how to orchestrate all this

@bindureddy Gemini even lack of creating simple comparison tables. It was before, they messed up lately

@bindureddy What could go wrong?

@bindureddy days? My calendar is marked. GPT-5.6 better be mind-blowing after all the teasing 😂

@bindureddy Simultaneous launches mean your fallback hierarchy ages out all at once. The bottleneck isn't picking the winner - it's running evals across 12 models before your prod traffic picks for you.

@bindureddy Let's see what happens with gemini 3.5 pro

@bindureddy What law allows the government to stop public releases of these?

@bindureddy But does the public get access to any of them?

What about people outside the US?

@bindureddy the synchronized launches are the leapfrog game in action. nobody wants to ship alone and hold the benchmark crown for a week before getting one-upped. launching into the noise is the strategy now

@bindureddy Brace for the ultimate compute crunch. 15 days until every major lab dumps model weights simultaneously. Infrastructure is going to melt, but open source dev speed is about to absolute skyrocket.