alt Hacker NewsGemini 3.8 Flash and 3.8 Flash Cyber
Comments
I've been using Gemini 3.7 for my personal trip planning app. Across multiple benchmarks, it ranks higher on everything I tried:
- Real world knowledge (when a thing opens and closes, the geographic region, historical facts). It's also the best at taking a cluster of places and working out a visiting order.
- Photo ranking (which photo should be the hero). Gemini can tell whether a photo is of the thing or of the view from it.
- Document parsing (extracting the relevant trip info from PDFs).
If you use LLMs for anything other than coding, I definitely recommend not discounting Gemini like I did just because other models are more popular.
Currently top at https://deepswe.datacurve.ai - beating Opus 5!
https://artificialanalysis.ai/models/gemini-3-8-flash shows an intelligence score of 59, the same as Opus 5 medium!
Wow - for a flash model this seems to benchmark powerfully. Remains to be seen what it is like to use.
Pelicans (thinking effort high, medium, low): https://tools.simonwillison.net/markdown-svg-renderer?url=ht... - high cost 8.9742 cents
Here are the 3.7 pelicans for comparison: https://tools.simonwillison.net/markdown-svg-renderer.html?u... - high cost 8.4387 cents
(I think thinking level low is a regression on 3.8 compared to 3.7.)
The most interesting thing about the Gemini models is still their multi-modal support: they accept audio and video input, OpenAI and Anthropic's flagships are still image-only.
Gemini Flash is also pretty cheap, so it's a great family for performing media analysis, like extracting structured data from images and video.
People have been sleeping on Gemini lately but these last few Flash releases (which were very rapid) are damn good.
These sort of fast and cheap models are great for tasks that are verifiable and can be retried infinitely (like coding), you can basically get frontier results with a good harness (at a fraction of the time and money).
I like Google's strategy here. These new Flash models of late (Flash 3.6, 3.7 and now 3.8) have obviously been distilled from a much larger unreleased model (Gemini 3.5 Pro, iirc from the rumors).
One aspect of model releases that don't get discussed as much are the cache invalidation (changes in underlying architecture, weights, or tokenizers); I assess Google seems to be squeezing the maximum out of the last 'Pro' version they released with 3.1 back in February.
Small models cataching up with their bigger siblings are fantastic news.
Wow this comes after what - 3 or 4 weeks since 3.7 Flash, which was also 3 or 4 weeks after 3.6 Flash IIRC?
I eagerly wait more info but sounds like Deepmind without Demis calling the shots has been unleashed and are operating at full speed? Shocker!
At this point it is a meme of course, but where is 3.5 Pro :)
Looks like the strategy of regular updates with incremental improvements is working out well. Interestingly, the biggest jump in Artificial Analysis Intelligence Index score is for reasoning level Medium ( 3.7 was 51, 53, 57 for Low, Medium and High, 3.8 is 52,57, 59 respectively). I think scores at lower reasoning levels are more indicative of model capability since higher reasoning levels are focussed on benchmaxxing. We use the lowest reasoning level in production with good results.
Wait, I didn't realize 3.7 Flash was already beating Sol on a bunch of the benchmarks. Isn't it a way smaller models?
I don't know if Google is having the worst marketing fumble or the most genius marketing one. Their "flash" models are very comparable to other companies' "pro" or "flagship" models. It seems to be a quite counterintuitive naming convention as it undersells the models.
Unless they have an even more powerful Gemini Pro in the oven...?
Something maybe unfamiliar with you: not about coding but writing. I've asked it to write an argumentative essay, which is a part of "gaokao" (China's university entrance exam), and its work is *extremely* impressive. speaks and writes like a real senior high school student, and the opinions unfold progressively with deep hierarchy. I don't know how the Gemini team reaches this because this kind of Chinese capability literally outperforms at least 2/3 Chinese students, no to mention those who speak Chinese. After all, the model speaks like a real humankind if you prompt it well. That's AGI guys
Also just want to let my appreciation here for 3.7 it’s cheap super fast super reliable incredible at information parsing eu host able (important for us) and perfectly integrated into gcp. Great job google!
Seem like a great, no-compromise, upgrade over 3.7 which is already a bargain, fast, and doesn't have the brain-damaged writing style of Claude.
"The knowledge cutoff date for Gemini 3.8 Flash is March 2026 – users can expect updated information for some domains while in others they may experience the model’s knowledge is limited to January 2025 (in line with the Gemini 3 Model Family)."
Kind of wild that they haven't (successfully) pretrained a base model since Jan-25.
The flash models, for coding are reckless in my experience. I have a Ultimate subscription, get good quota, but still use Opus 4.6 as it's much more reliable if you manage the context window carefully.
Is anyone here using using these models via google subscription (not api). I tried to in the past using gemini cli and then agy - headless invoked by codex and claude code, but they were so incredibly buggy that it stalled 1/2 times and I cancelled. Interested to know if that has changed!
Gemini 3.8 Flash is top of the Redactle LLM benchmark but so was Gemini 3.7 Flash. Both one shot all puzzles in the evals though 3.8 is just a bit faster. It also does the evals cheaper and faster than almost all the other models I've tried.
"available to trusted defenders through our new Fairwind Program"
3.7 flash was by far the best model for image recognition tasks according to my benchmarks. 3.8 flash didn't regress any candidates and improved some specificity (positive ID of common name vs species name of exotic fruit, correct identification of cast/replica of artifact and statue) but is still relatively weaker (26/30) on esoteric public figures (Korean beatboxers). I'm going to have to make my benchmark harder.
The iteration cycle is becoming very quick. Gemini 3.8 Flash arrived just 20 days after 3.7 Flash.
Similarly Qwen3.8-Max was updated in just 30 days (to the 0902 release) and Muse Spark in just 28 days (to the 1.3 release).
A year ago iterative releases were every 3-6 months. At what point will they reach nightly candidates?
The model seems fast enough to solve your problem before Google finishes explaining which of its three products you need to open to access it.
Still zero on the felony bench.
Is this weakness in their training regimen the impact of operating under regulatory frameworks for too long?
CAD for 3D printing is finally becoming feasible with Flash 3.7 and 3.8. Exciting times. https://github.com/ModelRift/openscad-skill/
As a big proponent of GPT-5.6-Luna for the combination of speed/perf/(especially)price,
Flash 3.8 seems like where I can specify Flash3.8 as the coding model as part of agent workflow.
The video recognition is especially impressive as they got all of Youtube to train from.
- Def people who has to queue video recognition jobs to use the model.
I have mixed feelings about Gemini 3.7 Flash. I used it for a personal project in Java and it was ok: it was crazy fast and it reached the correct result, but the code quality was barely passable.
I also used it for a an app for my Garmin watch, and it wasn't good. The code was compiling, but functionality was totally broken and even with a lot of steering it wasn't able to make it work. GLM 5.3-flash instead was up for it and the code wasn't bad at all. I am curious to see if 3.8 is an improvement in this use case.
I can’t wait until waiting hours and spending a big chunk of your usage per task seems antiquated, and real-time iteration on massive code changes is the norm. This might just be the year of efficiency, that truly allows AI to be used to the heart’s content.
How generous is the Google subscription quotas compared to Anthropic and OpenAI? This sounds like a really good potential model for high volume due to its speed and cost effectiveness.
(By high volume I mean things like "main app just updated with XYZ commits, please scan XYZ plugins and surface any compatibility issues")
It is more expensive per task than 5.6-sol high: https://artificialanalysis.ai/models/gemini-3-8-flash#price-...
Seems to do reasonably well in opencode according to artificial analysis. https://artificialanalysis.ai/agents/coding-agents
If I had to pay per token I would probably consider using this (they seem to be on the pareto of performance) but not being able to use opencode with a subscription is not really something I'm realistically going to do when claude and codex are around. Also never gotten along well with gemini-cli / antigravity-cli.
It seems like this is one of the most powerful models for the price, really didn't see that coming from Google
For my application, I'm still happily using gemini-2.5-flash and the only problem is when it reports being overloaded. It's for interpreting a downscaled phone camera photo of a hand-written shopping list on a whiteboard, and it works stunningly well. My handwriting sucks, too.
(I guess the only relevance here is that if your problem matches a model's strengths, then you can do fine with a model that is several generations out of date.)
Is the google infra stable enough right now? At the start of the year, the flash model was unusable for a whole month via gemini CLI. They could not fix it for a whole month and I was a paid customer.
There must be a deeper read on why Google can rapidly ship better small models while being delayed months on the bigger model.
What's the simplest explanation?
I think about Google is the value you get of their plans, for 5$ a month you get their ai plus model combined with 400gb you can share this with your family. The other ai companies don't provide family plans
I'm surprised the introductory 50% discount is good for 4 months. It seems like frontier models release new versions every 2-3 months, so raising prices in 4 months seems like a bad plan: you're effectively planning to charge users twice as much for a model that is no longer frontier.
Wow fable5.1 was the first model to do what I actually told it and I couldn’t find any problems with it, excited to try this just a day later lol
Still refuse to search internet for stuff it thinks does not exist lol.
And even when searching for internet, it still cannot suggest a up-to-date approach to the problem.
For example I'm using crystal, it recently revamped the concurrency/parallel model. Even using web search, gemini still does not aware of the new feature and still give the outdated code.
I'm sure my crystal usage is not the unique case here.
A company with 400+B revenue from software cannot build a usable command line cli for its vital AI model?
been absolutely loving 3.7 flash for coding. it feels very fast and quality is decent for implementing product features. usually use opus or sol for hardcore debugging.
Do they officially support you use their AI Pro subscription (or whatever the heck it's called this month, the one that gives you models in antigravity) in a 3rd party harness?
After struggling with Gemini for months, I think the trick to getting the most out of the model is writing a really solid personal intelligence/instructions prompt. The results are night and day in terms of performance.
It's interesting that Deepseek models were missing in the comparison. I see Deepseek v4 Flash a direct competitor to Gemini Flash for text-based agentic work.
So 89.4% on Terminal Bench 2 but only 19.1% on Tbench 4. Opus 5 is 89.1%/51.8%.
I don't use Gemini, but I thought `cool, let's give this new model a try`. Opened gemini.google.com, and I'm not even surprised. The drop down gives me the following options:
- Flash-Lite
- 3.6 Flash [new]
- 3.1 Pro
The above is why i don't use LLM products from Google. If the model is not available right this minute (heck, hours before the release!), then I'm not gonna bother getting back to it tomorrow, because tomorrow I'll be playing with the new model from OAI/Anthropic.
What is with Google's dumb ass STILL refusing to respect the OS dark mode setting in fucking 2027??
It’s strange that its score on Terminal‑Bench 4.0 is so low. They aren’t fast enough to benchmaxx that section.
The speed combined with the fact that this thing is really good at HTML JavaScript is pretty exciting.
Here's what I got for 1.8 cents and 13 seconds from the prompt "make me a cool thing in html":
https://gisthost.github.io/?6a77bc41a81718c6aaa10d4ab243c59f
Transcript here (it was part of a chat): https://gist.github.com/simonw/b6149a49d327164d67d62c3d12992...