we've added unique user rankings some models are token heavy so they skew upwards in rankings - unique people using the model is a more accurate ranking we'll orient more of our data around this metric

@opencode That makes a lot of sense. Focusing on unique users should give a much clearer picture of actual adoption rather than just raw usage volume. Go Deepseek!

@opencode Been using Deepseek v4 flash and mimo v2.5 , been real good for coding alot ngl. I haven't even used gemini pro and kept using these and server my purpose real good.

Still in beta, Polymarket already has best-in-class liquidity across virtually every category: pregame, futures, even in-play tennis. The depth is honestly absurd. Deposit $20, get free $50 with my code DOV50 Try trading on tonight’s game 👇

@opencode The only thing more discouraging than seeing Americans give their hard earned dollars and best ideas to adversarial models is the lack of US-based model orgs blowing them out of the water in response.

@opencode Can you use capital letters when you start a sentence?

@opencode This isn’t a good metric. This doesn’t tell which model is being used for serious work. I used GLM like once a day to maybe try a few things. Deepseek flash does so much work for me, same with pro model. In your chart all 3 models are the same. What is that supposed to tell me?

@opencode So this is how many unique users opencode has?

@opencode why so many people are using deepseek v4 flash ? for what kind of usecases people are using it ? is it intentional or is it the default choice in opencode ?

@opencode nice! I'm wondering if we can look into number of user's message using those models as well.

@opencode @opencode CLI... Cleanest interface there is. Bar none. @claudeai interface sucks but they have great models.

@opencode DeepSeek at the top is wild. Chinese Ai just keeps delivering on value while everyone else is still charging premium prices 😭

Rep Your Side. Make WC soccer picks in your state.

@opencode maybe consider weighting by the quota per month

@opencode I have a hermes agent connected to our CRM. Gave it 40 tasks (building a land sales app) and l got charged less than 3$ on open router using deepseek v4 flash. But yes l badly wanna buy a GPU

@opencode dsv4 pro have being my daily driver for personal projects, 🫶✨❤️

@opencode why can't I check how much token I spent:)

@opencode Guys, would be great if you measure tokens/session, as I think it can give us some idea how models are efficient If you separate it with session limits, then we can assume it was tokens/delivery

@opencode Really interesting. Shows the pure meritocracy in the open source llm market

@opencode Flash on top still, impressive?

@opencode GLM-5.2 is actually really good, honestly. Great price, great performance, and I've already integrated it into my product.

New here? Kick off with an exclusive welcome bonus just for you

@opencode Is bun issue on 1.3 version fixed?

@opencode Maybe a cost-token weighted metric? Old one heavy leaned to lightweight grunt work model like dsv4 flash and unique user is also kinda weird because some ppl mix use all models for different task🫥

@opencode are these analytics public? also can users review there as well?

@opencode unique users is better, but default model in ui still skews rankings hard

@opencode We want an IDE that doesn't overwrite code before review. Some code bases are to be delicate, and code shouldn't be overridden at prompt, something similar like Google Antigravity IDE A tool that gives software engineers control over "their code"

@opencode Wow cool, now can you please fix opencode so everytime i open a session it always loads the agents.md file, and add scroll bars to all panels so we can actually scroll since the direction buttons dont work?

@opencode If the scorecard counts people, I’m suddenly the top model, turns out my coffee‑fuelled queries finally count for something.

@opencode Can I see this data but just for the US?

@opencode So I will wait 23 days for my limit to reset? By then won’t my subscription expire?

@opencode so, I'm definitely using all 5 top models in the list.

@opencode GLM 5.2 is out ?!

@opencode Maybe request count is the correct answer here

Our delivery comes with a genius. Go learns what you like, builds your cart in seconds, and gets it all to your door in as fast as 15 minutes.

@opencode let deepseek cook

@opencode Rankings can be skewed by token-heavy use. Real cost comparison on oken.ai helps

@opencode 没有人能拒绝 deepseek-v4 的魔力！

@opencode Go girl. Kaya muyan. Basta dili pa manghapdos

@opencode Is here china....…