🚀 1,000+ TOKENS/S ON A 1T MODEL! 🚀
We are thrilled to release Xiaomi MiMo-V2.5-Pro-UltraSpeed in collaboration with @TileRT_AI , breaking the 1,000 tokens/s output speed on a 1 Trillion parameter model for the FIRST TIME!
Not wafer-scale integration like Cerebras. Not pure on-chip SRAM chips like Groq. We achieve 1,000 tps on a 1T MoE model using just a SINGLE, STANDARD 8-GPGPU NODE.
Read the full technical deep dive:https://t.co/MX0kjHKdKi
Want to experience the future of real-time AI?
👉 Apply for UltraSpeed now: https://t.co/aeWAxyhwVk
⏳ Limited-Time Access: Application-based · Jun 8 – Jun 23 (PDT)
💬 Chat Experience: Completely FREE for a limited time — try the blazing-fast web chat now.
⚡ UltraSpeed API: Just 3x the price for a ~10x boost in output experience.
🤝 Enterprise & Large-Scale Needs: [email protected]
@ism_sol Ga bisa bang pakai RAM, kayaknya emg harus pakai VRAM GPU. Saya pakai RTX2050 4GB ngejalanin starcoder atau model kecil yg 7B udah ngos ngosan sih. Coba cari tools namanya llmfit di github, bisa rekomenin model yg paling cocok buat machinemu
@kresbung stop nontonin video debat islam VS kristen dengan dalih memperkuat iman. Ada cara lebih baik memperkuat iman ketimbang nonton debat kusir agama yang emg didesain buat ragebaiting dan cari engagement, dan memuaskan ego belaka.
@InvertTheWing Mosquerra even got a yellow card for wasting time on the 46'. Literally one minute after half time. Yellow card. Wasting time. Minute one. Read that again.
Persetan perpanjang claude pro.
ada 82B token mimo gratis yang masih 97% nganggur.
reasoningnya surprisingly solid buat ukuran GRATIS.
Anthropic : "We're sorry to see you go!"
Gw : it's not you, it's the 82 billion FREE token 😭
PAKE INI BIAR CODING PAKE CODEX GA NGACO MULU
gua nemu bro custom instruction GRATIS buat Codex
buatan @karpathy
ini tuh AGENTS.md (plus instruction.md & SKILL.md) 🎯
gua taro aja Github repo nya di komen
tujuannya buat apa aja:
> bikin Codex jadi programmer senior yang super disiplin
> berpikir dulu sebelum mulai code
> gak over-engineer, kerjain persis yang diminta
> cuma ubah yang perlu, gak sentuh file lain
> sebelum bilang “done”, AI nya langsung cek sendiri targetnya udah bener
> ngurangin halusinasi, overbuilding & ngubah kode sembarangan
berguna poll kalo lu sering pake Codex buat coding (apalagi project gede atau production)
@ism_sol Aku dapat perplexity pro, free dari promo yg udah expired. Support banyak model premium, Claude sonnet, grok, GPT 5.5 dan berguna bgt buat ngerilis dokumen PRD, scraping dan research web, dan jadi buat bahan belajar. Kalau perplexity free rasanya unusable sih.