we've had intermittent issues with kimi 2.6 on OpenCode Go we're investigating with our providers, sorry about that hopefully can use one of the other models in the meantime

@opencode Is this on GO only? I added OpenCode Zen as an LLM provider for my Zen editor. I can see a lot of models, but not Kimi K2.6 (I see 2.5). Also, I see some messages about Deepseek, but I don't see it. I'm on Zen, not GO.

@opencode Yes. I had this problem a two days ago and switched to glm 5.1 since then

@opencode Opencode in terminal is editing files in plan mode. I had this issue with differents LLM's, and started to happend when i started with opencode go.

From deep questions to stunning visuals - Grok does it all. Sign up and start imagining instantly.

@opencode discovered deepseek v4 pro max because of this issue, not complaining

@opencode Yup just switched to Qwen 😂😂

@opencode Any updates on this? Maybe I missed it? Thanks for the transparency 🙏

@opencode Seems to be working fine now

@opencode We have 4 providers for kimi 2.6 llmgateway.io/models/kimi-k2… so when one falls, we route to the other automatically Try /connect and pick @llmgateway as the provider and paste your API key there. We also have similar Go plans here: code.llmgateway.io

@opencode Honest request you've done the 3x thing, but I used all my included usage before you started doing that, now I can't even try deepseek because it's not on Zen. Any relief for people like me?

@opencode Cant use most of the models right now. Not great. I paid for it.

@opencode I've been having issues with the WEB ui, ever since last night, it hangs, doesn't do anything just sits there, it was having issues connecting to the ghostly terminal. (restarted-updated-same issue). Sometimes does below in the picture, not sure if its even working. -nonexpanding

Stream. Scroll. Surf. Get speeds up to 400+ Mbps to connect with friends and family without delays.

@opencode My only problem is that I use kimi 2.6 too much and my tokens don’t last

@opencode I was wondering why my model randomly stopped.. glad its picked up!

@opencode I had this for last couplenof days. using deepseek instead

@opencode Thanks! I thought that it was only with me.

@opencode You guys should look into making open code available on Android and iOS.

@opencode This did prevent some work but luckily we have Deepseek V4 Pro

@opencode The app hangs a lot on windows, even worse when using DeepSeek

@opencode thanks for finally acknowledging it :)

@opencode We thank u for that because I discovered Deepkseek v4 and MiMo 2.5 Pro which are too goated

The books set the line. Our simulations tell you when they're wrong. MLB prop and game data built for bettors who actually do the work.

@opencode what is the provider? why vague posting

@opencode it affects all other resellers, the issue is with the provider you use

@opencode ran into similar issues. context window management is the rough edge, not the model itself. shorter focused sessions helped more than switching models

@opencode Reliability is the biggest hurdle for AI architectures right now. Multi-provider routing is definitely the right move to mitigate those server-side capacity limits. Redundancy is key for robust agentic workflows.

@opencode 3x for another weeks then

@opencode "Why does an AI model with a 1-million-token context window fail to support OpenCode and encounter errors specifically at the 250k token threshold when connected via a provider?"

@opencode provider incident messaging is underrated. naming the broken model and giving users the fallback path saves way more trust than pretending routing is invisible

@opencode Thanks for the update, switching to DeepSeek now.

BREAKING: This nurse blows the lid off of the Covid Shot. Now you know what happened in the hospitals as Covid happened!! Follow this account for more

@opencode Funny enough, the fallback model can bite you too if it lacks Kimi’s function‑calling support. Then you’re still stuck.

@opencode I paid the go plan, and i'm using minimax M2.5 Free, and it's amazingly good.

@opencode How I use this on my pc for free

@opencode kimi 2.6 is a great starting point, but the core problem isn't the model, it’s that providers are still deploying new models every 3 weeks with zero testing

@opencode Ah, Kimi 2.6’s been flaky for me too—especially on Go syntax highlighting. Tried DeepSeek-Coder as a fallback, works decently. Any ETA on the fix? 🤞

@opencode Just noticed. Thanks for the update.

@opencode is this resolved, and when grok?

@opencode redundancy tests like this validate the fallback strategy