๐Ÿ˜‚ Oh joy, another repackaged "innovation" โ€” the same old #software with a new twist of "RandomLabel" magic. ๐Ÿ™„ JavaScript's out, but fear not, you'll just need to sacrifice your #browser #privacy on the altar of Kimi K2.5! ๐Ÿš€
https://twitter.com/fynnso/status/2034706304875602030 #innovation #KimiK2.5 #HackerNews #ngated
Fynn (@fynnso) on X

was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID

X (formerly Twitter)
Fynn (@fynnso) on X

was messing with the OpenAI base URL in Cursor and caught this accounts/anysphere/models/kimi-k2p5-rl-0317-s515-fast so composer 2 is just Kimi K2.5 with RL at least rename the model ID

X (formerly Twitter)

Fili (@filiksyos)

Openclaw ํŒ: ๊ธฐ๋ณธ ๋ชจ๋ธ๋กœ 'kimi k2.5'๋ฅผ ์“ฐ๋ผ๋Š” ๊ถŒ์žฅ์ž…๋‹ˆ๋‹ค. ํŠธ์œ—์€ kimi k2.5๊ฐ€ Sonnet ์ˆ˜์ค€์˜ ์ง€๋Šฅ์„ ๋ณด์—ฌ์ฃผ๋ฉด์„œ ๋น„์šฉ์€ ์•ฝ 7.5๋ฐฐ ์ €๋ ดํ•˜๊ณ  Openclaw์—์„œ ๊ฐ€์žฅ ๋งŽ์ด ์“ฐ์ด๋Š” ๋ชจ๋ธ์ด๋ผ ์†Œ๊ฐœํ•˜๋ฉฐ, ์ปจํ…์ŠคํŠธ๊ฐ€ ๊ธธ์–ด์ง€๋ฉด ๋А๋ ค์ง€๊ณ  ์„ฑ๋Šฅ์ด ๋–จ์–ด์ง€๋ฏ€๋กœ /compact ๋ช…๋ น์„ ์‚ฌ์šฉํ•˜๋ผ๊ณ  ๊ถŒํ•ฉ๋‹ˆ๋‹ค.

https://x.com/filiksyos/status/2027732200612106341

#openclaw #kimik2.5 #llm #optimization

Fili (@filiksyos) on X

Openclaw pro tip Use kimi k2.5 as your default model Sonnet level intelligence but 7.5 times cheaper it's the most used model for openclaw As context grows, it will get slower and dumber so use /compact command

X (formerly Twitter)

Ivan Fioravanti แฏ… (@ivanfioravanti)

1์กฐ(1T) ํŒŒ๋ผ๋ฏธํ„ฐ๊ธ‰ ๋ชจ๋ธ(์˜ˆ: Kimi K2.5)์„ ๋กœ์ปฌ์—์„œ ์‹คํ–‰ํ•˜๋Š” ์‚ฌ๋ก€ ๋ณด๊ณ : ์ž‘์„ฑ์ž๋Š” ๋‘ ๋Œ€์˜ Mac Studio M3 Ultra(512GB)์—์„œ Apple MLX๋ฅผ ์‚ฌ์šฉํ•ด ์•ฝ 630GB RAM์œผ๋กœ ๋ชจ๋ธ์„ ๊ตฌ๋™ํ•ด ์ดˆ๋‹น 20ํ† ํฐ์„ ๋‹ฌ์„ฑํ–ˆ๊ณ , @exolabs์—์„œ ์‹คํ–‰ํ–ˆ์œผ๋ฉฐ @opencode๋ฅผ ํ™œ์šฉํ•ด ์ž๋™ ํ”Œ๋ ˆ์ด ๊ฐ€๋Šฅํ•œ ์Šค๋„ค์ดํฌ ๊ฒŒ์ž„์„ ์ƒ์„ฑํ•˜๋Š” ์‹œ์—ฐ์„ ์—…๋กœ๋“œํ–ˆ์Šต๋‹ˆ๋‹ค. ๋กœ์ปฌ LLM ์‹คํ–‰๊ณผ ์‹ค์‚ฌ์šฉ ๋ฐ๋ชจ๋ฅผ ๋ณด์—ฌ์ฃผ๋Š” ๊ธฐ์ˆ ์  ์„ฑ๊ณผ์ž…๋‹ˆ๋‹ค.

https://x.com/ivanfioravanti/status/2027278474155639133

#localllm #kimik2.5 #applemlx #macstudio #opencode

Ivan Fioravanti แฏ… (@ivanfioravanti) on X

Can we run locally a 1T parameters like Kimi K2.5? ๐Ÿ‘€ Yes we can! Here it is: - running at 20 toks/s on @exolabs with Apple MLX on my two Mac Studio M3 Ultra 512GB using ~630GB RAM - @opencode used to create a snake game with autoplay - You can see model creating the game and

X (formerly Twitter)

Islem Maboud (@Ipenywis)

Kimi_Moonshot์˜ 'kimi k2.5'๋ฅผ ์…€ํ”„ ํ˜ธ์ŠคํŒ… ๋ชจ๋ธ์œผ๋กœ ์ ๊ทน ์ถ”์ฒœํ•˜๋Š” ํ›„๊ธฐ. ์‚ฌ์šฉ์ž๊ฐ€ ์˜ค๋žœ ๊ธฐ๊ฐ„ ์‚ฌ์šฉํ•ด ๋ณธ ๊ฒฐ๊ณผ ์•ˆ์ •์ ์œผ๋กœ ์ž‘๋™ํ•œ๋‹ค๊ณ  ํ‰๊ฐ€ํ•˜๋ฉฐ, ์ด ๋ชจ๋ธ ๋•Œ๋ฌธ์— Anthropic ๊ตฌ๋…์„ ์ทจ์†Œํ•  ๊ทผ๊ฑฐ๋ฅผ ์ฐพ์•˜๋‹ค๊ณ  ๋ฐํž˜.

https://x.com/Ipenywis/status/2027126934954471801

#kimik2.5 #selfhosted #llm #model

Islem Maboud (@Ipenywis) on X

kimi k2.5 by @Kimi_Moonshot is king for self hosted models so many people are sleeping on it but I have been using it for a good amount and it just works I think I finally found an excuse to cancel my @AnthropicAI sub

X (formerly Twitter)

ใƒ“ใƒชใƒผใƒปใƒใ‚ฆ (@EikaChiu)

์›น ๊ฐœ๋ฐœ ์ฝ”๋”ฉ ์ž‘์—…์—์„œ ์ค‘๊ตญ์‚ฐ OSS ๋ชจ๋ธ์ด Top10์— 3๊ฐœ ์ง„์ž…ํ–ˆ๋‹ค๋Š” ์†Œ์‹์ž…๋‹ˆ๋‹ค. ์ˆœ์œ„๋Š” 6์œ„: glm-5(@Zai_org), 7์œ„: minimax-m2.5(@MiniMax_AI), 9์œ„: kimi-k2.5-thinking(@Kimi_Moonshot). ํ†ต๊ณ„ ์ถœ์ฒ˜๋Š” @arena์ด๋ฉฐ ์ค‘๊ตญ ์˜คํ”ˆ์†Œ์Šค ๋ชจ๋ธ์˜ ์•ฝ์ง„์„ ๊ฐ•์กฐํ•ฉ๋‹ˆ๋‹ค.

https://x.com/EikaChiu/status/2024038271773511743

#opensource #modelranking #glm5 #minimaxm2.5 #kimik2.5

ใƒ“ใƒชใƒผใƒปใƒใ‚ฆ (@EikaChiu) on X

Web้–‹็™บใฎใ‚ณใƒผใƒ‡ใ‚ฃใƒณใ‚ฐใ‚ฟใ‚นใ‚ฏใงใ€ไธญๅ›ฝใฎOSSใƒขใƒ‡ใƒซใŒใชใ‚“ใจ3ใคใ‚‚ใƒˆใƒƒใƒ—10ใซใƒฉใƒณใ‚ฏใ‚คใƒณ๏ผไปŠใฎๅ‹ขใ„ใ‚’ๆ„Ÿใ˜ใพใ™ใญใ€‚ 6ไฝ: glm-5 (@Zai_org) 7ไฝ: minimax-m2.5 (@MiniMax_AI) 9ไฝ: kimi-k2.5-thinking (@Kimi_Moonshot) Stats by @arena

X (formerly Twitter)

Alex Finn (@AlexFinn)

๊ณ„ํš ์ค‘์ธ ๋กœ์ปฌ AI ๊ตฌ์„ฑ ๋…ธํŠธ: Exolabs๋ฅผ ํ†ตํ•ด ๋‘ ๋Œ€์˜ Mac Studio์—์„œ Kimi K2.5๋ฅผ ๊ตฌ๋™ํ•˜๊ณ , ๋™์‹œ์— MiniMax 2.5๋„ ์šด์˜ํ•ฉ๋‹ˆ๋‹ค. Opus 4.6๋Š” 'Henry'๋ผ๋Š” ๋ฉ”์ธ ์˜ค์ผ€์ŠคํŠธ๋ ˆ์ดํ„ฐ๋ฅผ ๊ตฌ๋™ํ•˜๋ฉฐ, Mac Mini์—์„œ๋Š” ๋กœ์ปฌ MiniMax๊ฐ€ ์—์ด์ „ํŠธ๋ฅผ ๋‹ด๋‹นํ•˜๋Š” ๊ตฌ์กฐ๋กœ ๊ฐ ๋ชจ๋ธ์˜ ์—ญํ• ์„ ๋ถ„์‚ฐ์‹œํ‚จ ์„ค์ •์„ ์„ค๋ช…ํ•˜๊ณ  ์žˆ์Šต๋‹ˆ๋‹ค.

https://x.com/AlexFinn/status/2023608113530892786

#exolabs #kimik2.5 #minimax2.5 #opus4.6

Alex Finn (@AlexFinn) on X

Notes on my planned setup: Kimi K2.5 running across both Mac Studios using @exolabs Also will have MiniMax 2.5 running at the same time. Opus 4.6 will power Henry, the main orchestrator. The local MiniMax will power the agent on the Mac Mini. Kimi K2.5 will power the 2nd Studio

X (formerly Twitter)

Cerebras (@cerebras)

2018๋…„ BERT-Large(3.4์–ต ํŒŒ๋ผ๋ฏธํ„ฐ)์™€ ๋น„๊ตํ•ด ์˜ค๋Š˜๋‚  Kimi-k2 ๊ฐ™์€ ์ตœ์ฒจ๋‹จ ๋ชจ๋ธ์€ 1์กฐ ํŒŒ๋ผ๋ฏธํ„ฐ๋ฅผ ๋„˜๊ฒจ ์•ฝ 3,000๋ฐฐ ์„ฑ์žฅํ–ˆ์Šต๋‹ˆ๋‹ค. Cerebras์˜ ์—ฐ๊ตฌ ์ฑ…์ž„์ž @dmsobol์ด ํŠนํžˆ MoE(ํ˜ผํ•ฉ ์ „๋ฌธ๊ฐ€) ์•„ํ‚คํ…์ฒ˜๊ฐ€ ๊ทœ๋ชจ ํ™•๋Œ€ ์‹œ ์šด์šฉยท์‹คํ–‰ ์ธก๋ฉด์—์„œ ์™œ ๊ทผ๋ณธ์ ์œผ๋กœ ์–ด๋ ค์šด์ง€ ์„ค๋ช…ํ•ฉ๋‹ˆ๋‹ค.

https://x.com/cerebras/status/2023789540184584427

#modelscaling #moe #cerebras #kimik2

Cerebras (@cerebras) on X

In 2018, BERT-Large had 340 million parameters. Today, frontier models like Kimi-k2 exceed one trillion parameters, a 3,000x increase. In this conversation, @dmsobol, Head Research Scientist at Cerebras, explains why bigger models (especially MoEs) are fundamentally hard to run

X (formerly Twitter)

Bindu Reddy (@bindureddy)

์˜คํ”ˆ์†Œ์Šค๊ฐ€ ๊ฐ•์„ธ๋ผ๋Š” ๋‚ด์šฉ์œผ๋กœ Kimi K2์™€ GLM 5๊ฐ€ ์ƒ์œ„๊ถŒ์„ ์ฐจ์ง€ํ–ˆ๋‹ค๋Š” ์†Œ์‹์ž…๋‹ˆ๋‹ค. ๊ฒŒ์‹œ์ž๋Š” ์˜คํ”ˆ์†Œ์Šค์™€ ํด๋กœ์ฆˆ๋“œ ์†Œ์Šค LLM์„ ํ˜ผํ•ฉํ•ด ์ž‘์—…๋ณ„๋กœ ์ตœ์ ์˜ ๋ชจ๋ธ์„ ์ž๋™์œผ๋กœ ์„ ํƒํ•ด ํšจ์œจ์„ ๊ทน๋Œ€ํ™”ํ•œ๋‹ค๊ณ  ์•Œ๋ฆฌ๊ณ  ์žˆ์–ด, ๋ฉ€ํ‹ฐ-LLM ์ „๋žตยท๋ชจ๋ธ ์„ ํƒ ์ž๋™ํ™”์™€ ๊ด€๋ จ๋œ ๋ฐœํ‘œ๋กœ ํ•ด์„๋ฉ๋‹ˆ๋‹ค.

https://x.com/bindureddy/status/2022352056128868726

#opensource #kimik2 #glm5 #llm

Bindu Reddy (@bindureddy) on X

Open Source is ROARING - Kimi K2 and GLM 5 Take Top Spots Mix and match open-source with closed source LLMs We already do this for you automatically - maximize efficiency by choosing the best LLM for the task

X (formerly Twitter)

้‡‘ใฎใƒ‹ใƒฏใƒˆใƒช (@gosrum)

GLM-5์€ Claude Opus 4.5/4.6 ๋ฐ Kimiโ€‘K2.5์™€ ์„ฑ๋Šฅ์ƒ ํฐ ์ฐจ์ด๋Š” ์—†์ง€๋งŒ, ์ธํ”„๋ผ๊ฐ€ ์•ฝํ•ด ์ž‘์—… ์™„๋ฃŒ ์‹œ๊ฐ„(์‘๋‹ต ์†๋„) ๋ฉด์—์„œ ๋’ค์ฒ˜์ง€๋Š” ๊ฒƒ์œผ๋กœ ๋ณด์ž…๋‹ˆ๋‹ค. ์„ฑ๋Šฅ ์ž์ฒด๋Š” ์ค€์ˆ˜ํ•˜๋‚˜ ๋ฐฐํฌยท์ธํ”„๋ผ ํ™˜๊ฒฝ ์ฐจ์ด๋กœ ์‹ค์ œ ์‘๋‹ต ์ง€์—ฐ์ด ๋ฐœ์ƒํ•œ๋‹ค๋Š” ๊ด€์ฐฐ์ž…๋‹ˆ๋‹ค.

https://x.com/gosrum/status/2021792696080515572

#glm5 #claudeopus #kimik2.5 #llm #infrastructure

้‡‘ใฎใƒ‹ใƒฏใƒˆใƒช (@gosrum) on X

GLM-5ใฏClaude Opus 4.5/4.6ใ‚„Kimi-K2.5ใซๆ€ง่ƒฝ็š„ใซใฏๅŠฃใ‚‰ใชใ„ใ‘ใ‚Œใฉใ€ใ‚คใƒณใƒ•ใƒฉใŒๅผฑใ„ใŸใ‚ใ‹ใ‚ฟใ‚นใ‚ฏๅฎŒไบ†ใพใงใฎๆ™‚้–“ใงๅทฎใŒๅ‡บใฆใ„ใพใ™ใญ

X (formerly Twitter)