wafajan302
Region: PK
Tuesday 21 July 2026 14:12:54 GMT
11155
3072
312
63
Music
Download
Comments
Laila Khan 2345 :
nice
2026-07-21 20:03:56
1
Taj Muhammad :
jar❤️
2026-09-02 06:44:03
0
Fid@ wazir :
I miss you friend ❤️❤️❤️
2026-08-05 12:58:30
0
Ameen Ullah :
very nice
2026-07-30 03:35:14
1
اسماعیل جان :
2026-08-30 23:03:08
0
💎💍کیوٹ خانی💍💎 :
2026-07-28 20:50:25
1
Adi Khan :
good morning ko
2026-08-19 20:12:19
0
عرب تنھا :
2026-07-21 14:16:03
0
alidos :
🥰
2026-08-28 12:01:39
0
Moin Khan Dummar🚩 :
2026-08-19 07:26:34
0
yaseenkhanyasee978 :
🌹🌹🌹🌹
2026-07-29 18:53:22
1
Hanigul :
mashallah
2026-09-02 15:43:40
0
Ameen Ullah :
very beautiful
2026-07-30 03:35:08
0
To see more videos from user @wafajan..302, please go to the Tikwm
homepage.
![Replying to @aestheticbits CLAUDE SPEAKS PERFECT CLAUDISH 😭 Claude’s new models have a language of their own. Opus 5 averages around 510 words per answer in the benchmark you showed, while Fable 5 is still around 316. And yes you can prompt it to “be concise,” use a skill, or add system instructions. But that doesn’t necessarily solve the underlying problem. You’re now spending instructions/context telling an already verbose model not to be verbose, and depending on the setup, that can mean more token usage and more complexity just to get a short answer. For chatting? Annoying. For agents processing thousands of requests? That can actually matter. Do you like Claude’s detailed answers, or has the yapping gone too far? 😂 [Anthropic Claude, Claude AI, Opus 5, Opus 4.8, Opus 4.7, Fable 5, Fable 5.1, Claude models, AI models, LLM, AI agents, token usage, Claude API, AI coding]](/video/cover/7680728264578485535.webp)
