491–500 of 616 posts
My hunch is Google is trying to integrate a fast and relatively cheap AI across search and every other surface of their product suite. And for that objective, a model that can move faster while being accurate and cheap enough is more important to them than producing a frontier class heavyweight model.
Pelicans for 3.6 Flash and 3.5 Flash-Lite (Cyber isn't available to me through the API yet.) https://tools.simonwillison.net/markdown-svg-renderer#url=ht...
I am growing tired of these pelicans posts every time a new model is published. Feels to me like low effort personal brand promotion. Just sharing my 2 cents.
Earlier quoted context omitted.
Does it have "massive" margins? Afaik no one has said publicly what margins there are on an API call?
"As of October [2025], OpenAI's compute margins reached 70%, up from 52% at the end of 2024 and double the rate in January 2024, [The Information] said, citing a person familiar with the figures." https://www.bloomberg.com/news/articles/2025-12-21/openai-se... As for Anthropic, the rumors I remember seeing for their API margins were more like 85-90%, but I don't have a reference at hand for those. But once you know t…
Earlier quoted context omitted.
It seems like there are some credible rumors that Google is actually winning in terms of actually building models that work and don't lose money- between how they're able to price them, the TPU advantage and their capex advantage (being able to raise debt + just having a lot of cash - well I said not lose money... more like not go bankrupt). From the outside they look like they're behind in terms of frontier models,…
That "TPU advantage" might be slowing Google down (though likely not as much as their internal bureaucracy). Porting CUDA-based research, debugging, and overall experimentation speed is likely slower. The GPU is still king for training.
Earlier quoted context omitted.
It's rumored that Gemini 3.5 flash has a >50% margin, and I'd imagine 3.6 flash is even higher. I do not think OpenAI or Anthropic are actively chasing margins - though, Anthropic is supposed to be profitable on some form of non-GAAP accounting... I suspect Google isn't really interested in seeing how far it can get dragged into a race of selling dollars for $0.25, and is more interested to see if it can stay in the…
It kind of doesn't make sense though, because typically a large org like Google can afford to crush competitors on pricing. They could probably even go toe to toe with chinese model pricing for years without feeling it. Maybe they don't want to price war with the other labs so they can comfortably maintain healthy margins on selling them compute?
Earlier quoted context omitted.
Whats a better option for AA Coding Index?
DeepSWE and FrontierCode are more realistic if you read up on what they actually measure. But the most realistic is to try it yourself. Benchmarks can only vaguely represent typical usage, and how you judge the result. Giving the same real task you have to a few models will make you understand them better than chasing benchmarks.
I wonder how big the Pro model is that Google is using behind the scenes to train these smaller ones. Going on baseless speculation, the lack of accompanying pro models with these flash releases either means: 1) the model is too big to be economical, 2) google doesn't have the compute to serve the big model, 3) their big model has too many alignment issues to serve to the public. edit: looks like benchmarks are up on…
He says there are many similar vendors and teams with thousands of people in India and other countries.
Plugged 3.5 Flash Lite into an existing agent harness that was previously using 3.1 Flash Lite and this shit just does not work. It's not following instructions and is not producing the correct tool calls.
Maybe this is relevant? Just in case > For autonomous subagents with tool calls, code execution, or multi-step reasoning: set thinking_level to "medium" or "high" to prevent premature tool termination. I just happened to see that in the docs: https://ai.google.dev/gemini-api/docs/latest-model
Earlier quoted context omitted.
Yes agreed - I wrote this up a while back https://martinalderson.com/posts/whats-going-on-with-gemini/ My view then was they are optimising the models for inference ability on their own hardware AND use cases, which is often speed and time to first token. They've somehow seemed to end up with terrible compute shortages, which again is surprising given how good Google is at infra deployments AND have their own hardwar…
They include an LLM response with every single Google search, whether it is warranted or not. That scale is, my guess, many orders of magnitude higher than what OpenAI and Anthropic serve. And for Google none of these are paid interactions since their LLMs do not (YET) insert ads into the responses. So my guess is that Google will continue having compute shortages until the Gemini enshittification starts.
Let's assume Google serves AI overviews on every SERP (they don't) and don't cache them (they do, afiak).
And let's assume that each AI overview is 2000 tokens (blended input/output), that's 500T tokens a month.
It's rumoured that anthropic is serving somewhere close to 10Q tokens a month.
Now it may be that AI overviews uses vastly more tokens than that per search, but I doubt it based on speed to render the overview.
My very rough napkin math on this is that maybe AI overviews is consuming 100T tokens/month max (after adjusting for caching and SERPs that don't have them), which would be 1% of Anthropic token volume.
IMO Gemini has the best free tier models/app for everyday use. Muse-Spark is perhaps just slightly better, but has none of the connectivity to my GApps (for things like “create a recipe in my Google Docs from this image”). Plus they are probably running these things on every Google search so saving tokens is a huge win for them.
Free? Did I misread the pricing details?
Rankings for text are here https://arena.ai/leaderboard/text
For comparison of Free Tiers: - Gemini serves 3.6-flash (rank 12) - ChatGPT serves 5.5-Instant (rank 23) - Claude serves Sonnet 5 (rank 27) - Meta AI serves muse-spark-1.1 (rank 5)
While Meta AI serves the better ranked model, it doesn't end up working that well for other things. For example, if I ask "help me buy a new raincoat", it ends up suggesting a Cambodian website, whereas Google is well integrated with Google shopping. It doesn't have the same integration with GApps outside of Gmail/Calendar. A few other email connectors are available.
Claude has one of the best interfaces with connectors, skills, and plugins galore, but the model and limits are restrictive on the free tier.
Gemini, as far as I know, I've never hit a rate limit on Flash.
I believe Gemini is going to gain market share through the free tier funnel while serving models as cost-effectively as possible. People are going to use Gemini because they use GApps and Google.
ChatGPT and Anthropic are going to be competing for the API/Business users, but for everyone else they are going have to become Google before Google becomes them.