Trying Out OpenCode Go
Trying Out OpenCode Go
I’ve been using Codex until now, but I’ve recently subscribed to OpenCode Go, which is trending on X, and I’m trying out OpenCode.
I really liked Codex itself, but the only complaint was that ‘the responses were slow.’ Well, it’s just 3000 yen a month for ChatGPT Pro, so it’s understandable.
If you’re serious about using it, you’d probably have to pay as much as Claude.
At this time, the Go plan for OpenCode was trending, so I decided to try it out.
It seems quite usable for $10 a month (5 dollars for the first month).
The models available with the Go plan are as follows:
GLM-5.2
GLM-5.1
Kimi K2.7 Code
Kimi K2.6
MiMo-V2.5
MiMo-V2.5-Pro
MiniMax M3
MiniMax M2.7
Qwen3.7 Max
Qwen3.7 Plus
Qwen3.6 Plus
DeepSeek V4 Pro
DeepSeek V4 Flash
GLM-5.2 is said to be very smart, but I always feel like I’ll run out of tokens quickly if I use it, so I’ll stick with DeepSeek V4 Flash (Middle) for daily use.
I think I’ll switch to GLM or Kimi only when facing difficult problems.
Cost
By the way, the price per 1 million tokens is as follows, and DeepSeek V4 is overwhelmingly cheap.
Future of AI Coding
There are many talks about running multiple AI agents in parallel to create apps at incredible speeds, but I personally think that’s unrealistic.
How many tasks in actual work software can be fully entrusted to AI?
It’s not just a matter of trusting or not trusting AI. I think it would be quite scary to entrust systems that handle sensitive information, involve real money, or even life-and-death situations to AI.
Probably, most people will use AI to handle specific tasks while doing some design work themselves, gradually delegating coding tasks to AI.
You can request Claude or similar services to handle the design as well, but it becomes expensive. If you have plenty of money, that’s fine, but most companies can’t afford such luxurious coding.
By the way, there was someone on X who said that relying on expensive agents would be problematic in the long run, and I agree with that.
So I’ll keep using it for a while.
Comments
No comments yet.