@bindureddy Even the Deepseek v4 flash is giving me Opus 4.7 level of performance. 🔥🔥
English
Prajwal Ramteke
369 posts
Prajwal Ramteke
@Prajwal96343052
SDE-1 Frontend Engineer | 2.5+ yrs exp | Building scalable, fast UIs | React • Next.js • Remix • Node.js etc | GitHub: https://t.co/8K3AoQ5rvH
Nagpur Katılım Ocak 2021
90 Takip Edilen13 Takipçiler
DeepSeek V4 Beats Opus 4.7 And GPT 5.5 To Become The World's Best Open Source Model
DeepSeek V4 Pro is the NEW KING of open-source .
- better and 10x cheaper than Opus 4.7 and GPT 5.5 medium
- out performs Kimi 2.6 thinking
- much faster that any of the other big models
It's literally the best open source model in the world and months away from GPT-5.5 xHigh.
English
@YoussefHosni951 Just imagine token with deepseek v4 flash 🔥
English
Opencode x Deepseek v4 pro x High gives you 800 million tokens at only $10!
This is a solid option for personal projects at the Claude Opus 4.5 level.
You can go even better with Kimi K2.6 x opencode Go plan for $10/month. That's Opus 4.6 level
English
@ZypherHQ god mode with deepseek v4 flash github.com/PrajsRamteke/h… with my harness, if you want to you can try
English
I spent 1.7B tokens on Codex (Pro 5x) and 80M tokens on Claude Code (Max 20x) yesterday.
Guess which one gave me a usage-limit warning.
English
OpenCode Go is practically unlimited with Deepseek V4 Flash
Even at max thinking, which is pretty decent, I'm spending ~2% of my monthly limits a day
At that pace, it's ~50 8h session days a month!
Double that to 16h a day, and it gives you 33.3 days making ~4%/day its ceiling
Lotto@LottoLabs
Update on Opencode Go It’s great value for $5/month, there’s really no reason not to do the first month. At $10/month it’s still good value and gets you access to all sota OS models. You can’t daily drive it without hitting limits on the big models but w/ Kimi x3 you won’t hit limits unless you’re insane. Overall highly recommend the first month, then make your own decision.
English
I LOVE Deepseek-v4-flash, incredibly reliable and capable, logical. It's lacking in frontend but I have MiMo for that.
I would recommend any company spending 100k+ a year on AI to purchase 8-10~ 6000s and have a few of the works to have them blind test these models for work.
English
You won’t even imagine this:
I built my own harness and it refactored his own entire codebase in a single prompt using DeepSeek-V4-Flash.
Total token consumption: 11,441,820 🤯
One-shot refactor at scale.
crazyy output 🫡🫡👌
@deepseek_ai @opencode
English
@opencode there is issue with kimi k2.6 in opencode: Provider returned error i ma getting this issue for the kimi k2.6
i asked my harness agent (build by me, mode: deepseek v4 flash) to fix this, it changed the source code of opencode and fix the issue.
extremely crazy experience.🤯🤯🤯
English
Okay it's official, Kimi K2.6 3x usage on Go for another week
OpenCode@opencode
hmm should we keep kimi k2.6 at 3x usage for another week or nah?
English
@MountainMan246 @CodeByPoonam i tried with other harness, in Pi its works only not working in opencode
English
@Prajwal96343052 @CodeByPoonam Perhaps try again? I'm not having any trouble with it at the moment, but I read others are also having trouble.
Might be region specific errors?
English
Claude Opus 4.7 just got dethroned by a Chinese AI model.
And nobody's talking about it.
Kimi K2.6:
→ #1 coding model on OpenRouter
→ 7x cheaper than Claude Opus 4.7
→ 100 sub-agents in parallel
→ 1,500 total steps per task
Here's what happens when you build instead of hype
English
@kitlangton @opencode @zeddotdev Kimi k 2.6 is not working with opencode, in go plan @opencode help me
English
@CodeByPoonam opencode with the Go subscription ($5 first month) currently gets 3x the normal usage limits as promotion.
English
@MountainMan246 @CodeByPoonam kimi k2.6 is not working for me
i just sub to go plan and its not working
English
Just subscribe for the Kimi k-2.6 opencode go subscription, and I am getting this error. @opencode or can someone please help here
English
@cheatyyyy But Yes, Pi is a good harness boilerplate. you can modify it as you like and make it better than Claude, either.
English
@cheatyyyy I thing it's harness agent problem, I also face problem with opus 4.6 in pi harness, but try with Claude it done the task in under 5 seconds with sonnet 4.6 . I tried with my own Harness agent it's done the task under 10 second with sonnet 4.6. But pi take more that 2min.
English
