Sabitlenmiş Tweet
Introducing Response Caching: save tons of money and time on tests and agent retries.
Blog post: openrouter.ai/announcements/…
Available for free. Learn more 👇
English
OpenRouter
4.7K posts
OpenRouter
@OpenRouter
Discover and use the latest LLMs. 500+ models (incl. 50+ free), explorable data, private chat, & a unified API. https://t.co/qJG5mKrigL
Katılım Temmuz 2023
373 Takip Edilen94.5K Takipçiler
We analyzed GPT 5.5 vs GPT 5.4 and found that costs increased between 49-92%.
The 2x price hike of GPT 5.5 is mitigated by the model generating 19-34% fewer completion tokens for longer prompts.
More analysis here: openrouter.ai/announcements/…
English
@OpenRouter we love using OpenRouter with Frontman, been super useful for us from the beginning of implementing it. We're also using it with req_llm, also a great lib.
This is such a nice feature together with session tracking and all the other tracking features 🙏
github.com/frontman-ai/fr…
English
Introducing Response Caching: save tons of money and time on tests and agent retries.
Blog post: openrouter.ai/announcements/…
Available for free. Learn more 👇
English
OpenRouter retweetledi
I did it
I generated 100M tokens on @OpenRouter for free
All thanks to Owl-Alpha
For reference Opus 4.7 costs 5$/M input tokens and 25$/M output tokens
Let’s assume 100% of tokens generated were output tokens
Total cost for 100M tokens would’ve been 2500$
In 24hrs
English
OpenRouter retweetledi
@OpenRouter for devs this might not be useful, coz they can do this directly into their request handlers with redis, for vibe coders it make sense....
English
@YounesAka Yes! Works for all models. (But changing your model, of course, will cause a cache miss.)
English
Notice: there’s a bug in this Azure data that we’re working on correcting, due to a change in their error shapes upstream.
But gpt5.5 is seeing a lot of growth!
Matías@matidotlol
@theo 1. What
English
OpenRouter retweetledi
More about OpenRouter Enterprise: openrouter.ai/enterprise
It's best way to use LLMs at scale, and we're making it better every week!
mark@r_marked
@aritmiabattito No. Their pricing is extremely reasonable, and also gets better on Enterprise
English
@mylifcc of course! just use the actual resolved model slug that last worked for you
English
@OpenRouter semver works when 'minor' has a contract. but LLM 'minor' bumps freely change output. seen claude-3-5 to 3-7 quietly break a CLAUDE.md skill template after one update. alias resolved at request time, or pinned per-key? any way to roll back if a release breaks prompts mid-flight?
English
NEW: "-latest" model aliases 🔀
Route requests to "~anthropic/claude-opus-latest", "~openai/gpt-latest", etc to get the latest version of each major model. (Inspired by semver.) openrouter.ai/models?q=latest
Wes Winder@weswinder
@levelsio openrouter has a cool “nitro” flag in the model names to use the fastest provider so like “gpt-5.5:nitro” would be cool if the labs just let you use “latest” or something
English
Cache is scoped per API key, so different keys under the same account stay isolated. Cache hits don't count against provider rate limits since the request never reaches them.
In beta now. Docs: openrouter.ai/docs/guides/fe…
English
Controls:
1. `X-OpenRouter-Cache-TTL` to set your lifetime (1 second to 24 hours, defaults to 5 min)
2. `X-OpenRouter-Cache-Clear` to bust a specific entry Response headers (HIT/MISS, Age, TTL) so you can see what's happening
Or set cache_enabled: true on a preset and forget the headers.
English
OpenRouter retweetledi
S/o @OpenRouter for building the best AI routing solution on the market. If you are working with LLM APIs, you should be using them, full stop.
English