It was fun while it lasted, but I see Luna is back up to $1.20.
At half price for me it was hands down the best value model for coding.
Honestly it's probably still a good deal even at that price, but I'm wondering for coding at that price does it still hold the top spot, or are you guys using something else?
Hi,
I'm usually not too bad at common bird calls, but the other day I heard a bird repeating a phrase that I didn't recognise. My best guess is Great Tit as they can be quite varied, or maybe Coal Tit but I'm not sure. I didn't get a look at all at the bird.
I've included the sonograph as well to indicate the phrase. Its three quick notes and then a descending note. It was repeating this infrequently with no variation to the phrase.
https://reddit.com/link/1vq608w/video/z052a1h2fsjh1/player
Thanks!
I was very excited to to see that GPT5.6 Luna was only $0.1 IN / $0.60 out including both OpenAI and OpenRouters discounts together.
Wow!
So I had a play around today, and this is the PR for reference: https://github.com/SuperToolMake/supertoolmake/pull/149
The cost shown in opencode was $1.35, and after digging through the input, output, reasoning, cache pricing, etc I was able to work out that the price was correct. A bit convoluted but fair as all of those prices are listed.
To my surprise OpenRouter was showing $2.30 usage in the dashboard. I was concerned that it was a bug, or that the OpenRouter 50% off wasn't being applied.
After digging around the dashboard in OpenRouter, I couldn't find a specific breakdown of the cost per token type, so I asked Luna to investigate, and got this response:
>The discrepancy is real and appears to be caused by OpenRouter’s long-context pricing.
>OpenCode calculates cost locally from its cached model pricing:
>- Input: $0.10/M
- Output: $0.60/M
- Cache read: $0.01/M
- Cache write: $0.125/M
>OpenRouter applies a higher rate when the prompt exceeds 272K tokens:
>- Input: $0.20/M
- Output: $0.90/M
- Cache read: $0.02/M
- Cache write: $0.25/M
>This task had many large-context requests:
>- Approximately 221 requests exceeded the 272K threshold.
- Approximately 118.5M cache-read tokens were used.
- The higher cache-read rate alone adds roughly $0.84.
- Recalculating with OpenRouter’s long-context rates gives approximately $2.297, matching the $2.30 shown by OpenRouter.
>So the numbers represent different things:
>- OpenCode: local estimate based on its cached base pricing, around $1.35-$1.37.
- OpenRouter: provider-side billing using the applicable long-context rates, around $2.30.
>OpenRouter’s usage or generation details are the billing source of truth. The OpenCode UI does not currently appear to account for OpenRouter’s conditional long-context pricing override. The PR metrics section contains the OpenCode estimate, not the authoritative provider charge.
It's a bit silly, and ironic, that I had to use an LLM to figure out the OpenRouter pricing.
Also it's a bit weird to have a 50% off sale, but then nearly 50% of the cost was hidden.
Anyway, assuming this information is correct, these prices should definitely be listed within the OpenRouter dashboard beside the models. (If it already is, please do point me in the right direction).
OpenRouter is a fantastic service, and for $2.30 I'm still happy I got a good deal for the work done, but no one likes hidden charges.
Please make the long context pricing more transparent. Show it on the model page alongside the input/output pricing.
Complete crypto noob here, but I wanted to try out buying credits with USDC.
I have some USDC on Coinbase.
On OpenRouter, I do Buy Credits -> Use crypto.
I see 'Select a wallet' but how do I pay with my Coinbase USDC? I don't see a Coinbase option. I don't want to use a phone app, just the browser if that's possible?
Hi everyone,
I'm currently working on the LarusDB project. The idea is to create a database of Larus (large gull) species, labelled with age, location and observation date.
Gulls, especially Larus, can be notoriously difficult to ID, so I'm hoping this can be used as a learning tool for newer "gullers" while also being a place to see regional differences of the same species.
There's currently over 1500 observations, with many of those being my own, but others have kindly given me permission to upload their photos as well.
I am looking for additional contributions, if anyone has a collection of images they'd like to share on Flickr or similar, it would be very much appreciated.
If you don't have any images to share, but know someone that might be interested in the project, please share this post with them!
Currently it's more European focused, but I'm open to including Larus species from around the world.
Here is a link to the app: https://app.gullinfo.org/app/larus/#/gull
If you are interested or have any feedback, please do drop a message below!
I will add a reference to your name / photo collection by default, so let me know if you'd rather contribute anonymously.
Hey,
Just wondering if it's possible to capture a Dragonfly in-flight with a bridge camera? Might be asking a bit much here, maybe recording a video is the best bet.
Here's a couple of photos I managed to get (I think of a male Common Hawker) with 1/1000 shutter speed, continuous auto-focus (which struggled a lot - they're so fast!).
Hi everyone,
I'm currently working on the LarusDB project. The idea is to create a database of Larus (large gull) species, labelled with age, location and observation date.
Gulls, especially Larus, can be notoriously difficult to ID, so I'm hoping this can be used as a learning tool for newer "gullers" while also being a place to see regional differences of the same species.
There's currently over 1500 observations, with many of those being my own, but others have kindly given me permission to upload their photos as well.
I am looking for additional contributions, if anyone has a collection of images they'd like to share on Flickr or similar, it would be very much appreciated.
Currently it's more European focused, but I'm open to including Larus species from around the world.
Here is a link to the app: https://app.gullinfo.org/app/larus/#/gull
If you are interested or have any feedback, please do drop a message below!