Sabitlenmiş Tweet
Mateusz Mirkowski
3.8K posts
Mateusz Mirkowski
@llmdevguy
Autonomous agents, agentic engineering Building & testing agentic systems Exploring local LLMs
Remote work evangelist Katılım Mart 2013
150 Takip Edilen1.8K Takipçiler
We have it!
Looks fantastic.
MiniMax (official)@MiniMax_AI
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities - Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas - MiniMax Sparse Attention scales context to 1M - Natively Multimodal from Step Zero API: platform.minimax.io Token Plan: platform.minimax.io/subscribe/toke… 🚀New! MiniMax Code: code.minimax.io Weights & Tech Report in ~10 Days
English
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities
- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero
API: platform.minimax.io
Token Plan: platform.minimax.io/subscribe/toke…
🚀New! MiniMax Code: code.minimax.io
Weights & Tech Report in ~10 Days
English
Playing with new TV(middle one).
My wife loves the new bedroom design for the next two weeks - 10/10.
My wife happiness - 10/10
My happiness - 10/10.
Only one score is true. Guess which one. 😂
BTW new soundbar is coming on Monday. 🙈
English
@llmdevguy Bro, I don’t like anthropic, their way to treat individuals sucks, but you have not choice if you’re not a coder, for coding, just throw Claude out, codex is good enough
English
Did Opus 4.8 beat GPT-5.5?
Please tell me because I don't use anything from Anthropic.
English
@llmdevguy I've literally abandoned Anthropic.
Everytime that I've tried to back to Opus, its a disaster. Then, i have to use code to fix all the findings.
English
@llmdevguy Opus 4.8 beat GPT-5.5... at burning through my tokens faster.
English
@llmdevguy i prefer gpt 5.5 over opus 4.7. however, after some testing over the past 2-3 hours i find 4.8 to outperform 5.5 in most of my tasks.
English
After spending time with Qwen 3.6 27B in Cursor I’ve come to realise the constraint on local models isn’t intelligence but the harnesses.
Local model harnesses are TERRIBLE.
Pi, open code etc are genuinely bad.
As a community, we need to do better than this.
English
🇨🇳Xiaomi increased mimo usage.
This is DeepSeek effect.
Most companies reduce usage.
Deepseek and xiaomi increase or reduce price.
Win for us.
Xiaomi MiMo@XiaomiMiMo
🚀 Better inference efficiency, lower costs, broader access. MiMo-V2.5 Series API pricing is now permanently reduced — by up to 99% compared to previous pricing. ✨ Unified pricing across all context lengths. MiMo Token Plans have also been upgraded: • 5–8× more usable tokens at the same price • Simpler and more transparent billing rules 🎁 As a thank-you to current users, all current Token Plan credits will be fully reset. 🎧 MiMo-V2.5-TTS remains free for a limited time. ⏰ Effective May 26 at 6:00 PM PDT. These improvements are powered by continued inference optimization and serving efficiency upgrades across the MiMo stack. 🛠️ We’ll also publish a detailed technical blog on the inference optimizations later — stay tuned.
English
Official account for Supergemma launched.
Let’s keep building 💪
SuperGemma@0xSUPERGEMMA
This is official account of $SUPERGEMMA We build ecosystem for open-source developers. Dev: @jun_song Open source will win.
English
Unbelievable.
For this price it's amazing choice for coding and Hermes agent.
DeepSeek@deepseek_ai
We are making our discount permanent! 🎉 Enjoy building with DeepSeek-V4-Pro and bring your innovative ideas to life! 🚀
English
8gb vram for local models?
Still usable.
You don't have to start with rtx 6000 pro.
witcheer@witcheer
running Hermes locally with Qwen 3.6-35b-a3b is possible on a RTX 4060 Ti 8GB. my params are: ~~~ llama-server \ -m ~/llama-models/Qwen3.6-35B-A3B-UD-Q4_K_M.gguf \ -ngl 999 -ncmoe 30 -fa on \ --cache-type-k q8_0 --cache-type-v q8_0 \ -c 32768 -n 8192 -np 1 -t 6 \ --reasoning off \ --no-cache-prompt --checkpoint-every-n-tokens -1 \ --jinja --metrics --host 0.0.0.0 --port 8080 ~~~ biggest flaws: - context: if you are coding, a few prompts will eat it all - speed: it took 17min to create a medium-difficult .py file but it works! I'm going to test /goal feature as well, to see how Qwen handle multiple compactions and see if it can finish a goal.
English