frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Gemini 3.7 Flash

https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/
44•meetpateltech•56m ago

Comments

spelk•50m ago
>3.7 Flash is available through the end of the year at an introductory price 1 of $0.75/1M input tokens and $3.75/1M output tokens. This price combined with the enhanced model performance enables developers and customers to scale production-ready agents cost effectively.

Introductory pricing until December 2026 implies no significant Gemini Flash developments until the next year.

randomblock1•26m ago
I think it's just meant to make it more competitive, Gemini has kinda been behind in everything except maybe multimodal. It's only 3 weeks after Flash 3.6, so if they really wanted to, they could probably do a 3.8 Flash before then.
nateb2022•21m ago
Or a 3.7 Flash-Lite
re-thc•21m ago
> implies no significant Gemini Flash developments until the next year.

Gemini 4 is apparently just around the corner so unless there's a 3 month delay... there's at least a new Flash update.

bisonbear•33m ago
They compare it to 5.6 Terra, however https://cognition.com/frontiercode puts Terra at about 1/2 the price

Also have to compare to the recent Grok 4.6 release, which appears to straight up be better AND cheaper

Hard to understand why anyone would choose 3.7 Flash under these conditions.. is Deepmind still a frontier lab?

ValentineC•26m ago
> Hard to understand why anyone would choose 3.7 Flash under these conditions.. is Deepmind still a frontier lab?

At this point, I think they're mostly targeting Google One and Workspace subscribers, except doing worse compared to Microsoft because they don't have Microsoft's huge enterprise moat built from their DOS and Windows days.

cahaya•12m ago
Agree, with you but I'm still using 3.6 Flash because of tok/s/ latency/ uptime with high context. Tried Grok 4.6 and it was scoring lower on some internal benchmarks or slower.
mdasen•10m ago
Artificial Analysis shows Grok 4.6 taking $1,068 to run their suite while Gemini 3.7 Flash takes $485. So it looks like Gemini 3.7 Flash is less than half the price in the real world.

Per-token cost isn't a great metric given that some use way more tokens than others.

nickandbro•30m ago
This is genuinely a competitive model, considering it beats Claude Sonnet 5 on almost all benchmarks and is more than half its price. Seems like Google is back in the game, though not leading the frontier anymore.
9cb14c1ec0•27m ago
Claude Sonnet 5 is such a garbage model, so not sure what that says about Google's new best model.
onlyrealcuzzo•27m ago
Sonnet 5 is arguably the most cost ineffective model to ever be released, so that's not really impressive.

It can regularly cost more than Fable, take longer, and deliver far far lower quality.

I'm much more interested how this compares to Luna - which on price is terribly - but at least on quality the benchmarks make this look competitive / usable.

If Google continues monthly Flash releases like Sundar said they would, and they continue to have this much of an improvement in cost/quality - then in a few months this could reasonably be very competitive with the best of the best.

It is not there yet, but at least it's super fast, I guess.

nickandbro•25m ago
Agreed
qeternity•8m ago
> more than half its price

Less than half its price.

More than 50% discount.

jespinel•3m ago
IMO, they should drop their previous model (3.6 Flash) from the benchmark charts. I don't care how better this is compared with their previous model. What matters (to me) is:

1. How the new model performs against the other top models in the same category.

2. The pricing of the new model against the other top models in the same category.

How Compaction Works in Pi

https://earendil.com/posts/compaction-in-pi/
1•tosh•32s ago•0 comments

Shadows: Sample app that demonstrates techniques for realtime shadow mapping

https://github.com/TheRealMJP/Shadows
1•klaussilveira•3m ago•0 comments

Enable P2P PCI transfers on Nvidia 3090, 4090, 5090

https://github.com/aikitoria/open-gpu-kernel-modules
1•jacquesm•4m ago•0 comments

Video: I'm Done Coding with AI

https://www.youtube.com/watch?v=2ZU3j4GQ4K8
1•champagnepapi•4m ago•0 comments

Solid 2.0 RC: The Big <Reveal>

https://www.solidjs.com/blog/solid-2-0-rc-the-big-reveal
1•GavinAnderegg•5m ago•0 comments

Some Virtues of Narrative Poetry

https://newversereview.substack.com/p/the-purpose-of-poetry-is-to-tell
2•samclemens•5m ago•0 comments

Previewing Ultrafast mode: GPT-5.6 Sol at up to 14x the speed

https://twitter.com/openai/status/2087947721936359705
2•qb•7m ago•0 comments

Show HN: Tired of hand-writing fake API data, so I built a CLI for it

https://github.com/SukhdevThukral/mockit
2•sukhdevth_•7m ago•0 comments

Where did the old web go? We followed 657,607 links to find out

https://0.mk/blog/link-rot
3•tdx•9m ago•0 comments

Namecheap phoenix datacenter overheats taking all sites down

https://www.pcmag.com/news/is-your-website-down-outage-hits-hosting-provider-namecheap
3•rgbrgb•9m ago•0 comments

New model BDH-CQ costs $0.007 per task 11x less than OpenAI Luna even w 80% off

https://huggingface.co/papers/2608.09888
3•wordpad•9m ago•0 comments

Choose Boring Technology (2015)

https://mcfunley.com/choose-boring-technology
7•tosh•10m ago•0 comments

Gemini 3.7 Flash (High) Intelligence, Performance and Price Analysis

https://artificialanalysis.ai/models/gemini-3-7-flash
4•theanonymousone•10m ago•0 comments

Show HN: Sued-rs – a fake demonic oracle for the terminal, in Rust

https://github.com/Danilo-Guedes/sued-rs
2•rocknbrain•11m ago•0 comments

SecurePkt – High-Speed, Post-Quantum UDP Tunnel Protocol and SOCKS5 Gateway`

https://pypi.org/project/securepkt/
2•mohammedqannan•11m ago•0 comments

Donkey.bas is 45 Years Old – 131 line of Glory

https://donkeybas.com/
7•jkrauska•12m ago•2 comments

Show HN: Taurus Agents, my take on multi-agent hierarchies

https://taurusagents.com/
2•sergevar•12m ago•0 comments

A Little Introduction to Control Flow Integrity - James McNellis - C++Now 2026

https://www.youtube.com/watch?v=FNHMwi_0psQ
2•edward28•13m ago•0 comments

SaaS Onboarding Benchmarks Report 2026 (real data from 464 products)

https://produktly.com/research/saas-onboarding-benchmarks-2026
2•Kkoala•14m ago•0 comments

Is OpenAI hardware, the Apple of tommorow?

https://s-1.vercel.app/posts/the-struggle-of-openai/
2•Taikhoom10•14m ago•2 comments

Flock (Again) Activates a Camera System a Town Had Voted to Shut Down

https://www.techdirt.com/2026/08/13/flock-again-activates-a-camera-system-a-town-had-voted-to-shu...
3•hn_acker•15m ago•0 comments

X: Open-Sourcing the for You Timeline

https://twitter.com/XOpenSource/status/2087951962004230428
3•tosh•16m ago•0 comments

Additional publications in mathematics by Jeremy L. Martin

https://jeremymartinmath.github.io/morepubs.html
1•fanf2•16m ago•0 comments

New York Wants to Make Use of Its Oppressive Subway Heat

https://www.governing.com/transportation/new-york-wants-to-make-use-of-its-oppressive-subway-heat
2•fkozlowski•17m ago•0 comments

We spent ten years burying the 10x engineer myth. AI is digging it back up

https://maxence.maireaux.fr/
2•flemzord•18m ago•1 comments

Instagram's New Logo

https://www.theverge.com/tech/979583/this-is-instagrams-new-logo
3•MehrdadKhnzd•18m ago•0 comments

Compute-Optimal Is Not Cluster-Optimal

https://szha.ai/blog/compute-optimal-is-not-cluster-optimal/
1•yuxin_tang•18m ago•1 comments

Modeling One LLM Agent Three Ways: Python, Clojure, Elixir

https://sdtimes.com/programming-languages/elixir-clojure-or-python-for-llm-agents-our-experience-...
1•ViktoriiaYarosh•19m ago•0 comments

Ask HN: What's slop? what's AI written text and why read/not read?

2•xlayn•20m ago•2 comments

Hooray for index funds–just don't call them passive

https://www.economist.com/finance-and-economics/2026/08/11/hooray-for-index-funds-just-dont-call-...
2•thm•20m ago•0 comments