Mateusz Mirkowski (@llmdevguy) header
Mateusz Mirkowski (@llmdevguy) profile picture

Autonomous agents, agentic engineering Building & testing agentic systems Exploring local and Chinese LLMs

Remote work evangelist
Joined March 2013
Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)Photo by Mateusz Mirkowski (@llmdevguy)
Free Qwen4-Max for Hermes. Great addition to free Hy4. Free models == best models. 馃槃
Qwen4-Max, the new flagship model from Alibaba, is free in Nous Portal for 5 days in partnership with @inferact and @Alibaba_Qwen. It鈥檚 built for long-horizon work, with native vision and 1M context. Try it today at portal.nousresearch.com
2
182
Replying to @MicWasowski
Niech Jensen wr贸ci. 馃檹
22
Replying to @bindureddy
Deepseek, qwen 3.8 and next flash are kings.
56
That's true. Google models feels different. Shame I don't have access to Carbon. :(
1
25
Replying to @loktar00
I like all of them.
1
14
Replying to @7AdamK
We know that.
373
So Astra 6.2 sucks. Oh wait maybe not? 馃槃
I tested GPT-6.2 Astra on real work. 2 repos, 105 replanted bugs. Find and fix what you can. It's the new #1. The results: - GPT-6.2 Astra (max): 62.7 for $28.35 - GPT-6.2 Astra (xhigh): 60 for $23.53 - Fable 5.2 (max): 58 for $139.65 - Gemini 4 Carbon (high): 57.3 for $31.32 - Sonnet 5.5 (max): 51.3 for $154.03 - GPT-6 Astra (max): 45 for $33.03 From GPT-6 Astra to GPT-6.2 Astra: +17.7 points at max, +17 at xhigh. Whatever people say about other use cases, it's a code review beast. GPT-6.2 Sol and more effort levels drop soon in the thread 馃У
1
2
214
Replying to @theo
Ok but what about Sol?
197
Replying to @Hibara_AI_lover
It was very generic prompt like create house with hd-2d style.
12
Of course these numbers are fake, but latest googles models are really good. I hope new google model will have numbers like neon. 馃グ
Ok Google is back. That鈥檚 official.
152
Replying to @jeSIs_x @Teknium
Hermes is still the king. 馃憫
1
140
Replying to @kapilansh_twt
This is so annoying. I often asked something Claude in my language and he often respond in English.
1,003
Replying to @thatoneguy_78s
We will see soon. :)
2
997
Replying to @theo
It鈥檚 time to switch to Omarchy.
15
This is true strength of Sol - amazing code review and security checks.
1
1
85
Replying to @daveeyyy_
Of course it is.
66
Replying to @ZainanZhou
Cheaper and better.
1
39
Replying to @stevibe
I recommend to use Gemini for specs to safe limits in antigravity.
1
2
36
Ok Google is back. That鈥檚 official.
gemini 4 neon numbers floating around
10
8
1
410
40,520
Replying to @DanielPetroAI
Me too. With deepseek 4.1 flash its almost feels unlimited.
21