Jane Doe

@BitgetWalletCN @0xAA_Science 肯定不，换行还需要 \n token

@0xAA_Science 难道 这样 比较 省 Token 吗？

ChatGPT 每行只蹦几个词的说话风格，是哪个傻逼训练的？

@noor36758 @jahirsheikh8 Five? How? I can only see three and four

@jahirsheikh8 Cutting a corner off a rectangle can give you 3, 4, or 5 corners depending on the cut... AI answered confidently, geometry answered correctly

This is exactly why AI won’t be replacing mathematicians anytime soon.

@thebadmaash @anyelamarillo There's a 250,000 tokens per minute limit

@janedoe2ooo @anyelamarillo So token counts don't matter?? Give me estimate in tokens.

🚨 GEMINI 3.5 FLASH YA ES GRATIS 😱 Google lo activó en silencio. Mientras la mayoría sigue hablando de los viejos modelos 2.5, tú ya puedes usar el modelo más nuevo de Google sin pagar ni un peso. Lo que obtienes por $0: • Gemini 3.5 Flash (el más potente actual) • Gemini 3.1 Flash-Lite (perfecto para alto volumen) • 1 millón de tokens de contexto • 1.500 solicitudes/día • Multimodal completo (texto, imagen, audio y video) • Compatible con OpenAI (solo cambias la URL) Modelo de frontera + 1M contexto + gratis. Mientras la mayoría de guías todavía recomiendan modelos antiguos. Guarda este post 🔥 El nivel gratuito cambia casi todos los meses. Configuración completa paso a paso 👇 (en comentarios)

@BytePlusGlobal I tried, to be honest it's better than gpt image 2. The lines are cleaner and straighter, and there's no weird color error. Good job

What makes Dola Seedream 5.0 Pro (hereafter referred to as Seedream 5.0 Pro) different? It starts with control. Let’s start with one of the biggest shifts in AI image generation: Interactive Precise Editing. Instead of repeatedly rewriting prompts, edit exactly where changes should happen while preserving the rest of the image. From product edits and sketch-guided creation to color, material, and multi-image editing, Seedream 5.0 Pro gives creative teams greater precision throughout the design workflow.

@haider1 Good, I don't like Geminis tone, a new pretrained model would be welcomed

if gemini 3.5 pro is just 3.1 pro with better instruction following and stronger Antigravity skills i'll take it but multiple rumours suggest it's a newly pre-trained model with major changes to its architecture, data, and training process not simply an improved 3.1 pro

@thebadmaash @anyelamarillo A lot. If you use it for eight hours straight, you can send up to approximately 3.2 messages per minute

@anyelamarillo 1500 requests is how much usage?

@Neolie_e26 @CodeWithAmann 3.5 Flash is actually specifically trained for agentic use cases, bud that sucked

@janedoe2ooo @CodeWithAmann À part pour les modèles de Google qui ne sont pas développés pour cela spécifiquement

Be honest, which Ai coding tool has the strongest future

@ChinaEastern_CN Prove it tho you dont have to

谁来救救我…… 最近总有人对我说： “别再装日本人了。” “别冒充日本人！” 之类的话。 我确实曾无数次希望自己能出生在中国🇨🇳，但无论我怎么想，我的国籍始终都是日本，我也是一个真正的日本人。 拜托了，不要再这样攻击我了……

@ai_xiaomu Opus 5? 你是说 Sonnet 5?

7月才过去10天，已经卷成这样了： • Fable 5：有史以来最智能的AI模型 • Muse Spark 1.1：Meta重返赛场 • Opus 5：Fable的平价版本 • Grok 4.5：Opus的平价版本 + Elon回归 • GPT 5.6：平价、闪电般快速、在某些方面优于Fable • GPT 6：Goblin模式 • GPT Work：令人惊叹的全新工作方式 难以想象年底或者明年会怎样🤣

@Monika2r @CodeWithAmann @grok It has free image gen tho better than meta lol

@CodeWithAmann @grok Copilot

Hey @grok remove the worst ai from these

@nice_img @corndogjpn2 两小时说实话够慷慨了

@corndogjpn2 plus用户可以使用2小时，我已经用完了

你们快放下手里的事情，打开自己的手机的ChatGPT客户端，看一下是不是已经更新到GPT Live这个模型。 和他对话1分钟，做一个非常差的交谈者，不停地打断，转换话题，插话，recall，它完全没有问题。 我刚才和他对话了3分钟，我现在已经瘫坐在椅子上，仿佛看到了原子弹爆炸

@ItsAlexhere0 baguette

USA has ChatGPT USA has Grok USA has Claude USA has Gemini USA has Llama USA has Copilot China has DeepSeek China has Qwen China has Ernie China has GLM China has Kimi China has MiniMax Europe has?

@TimJayas In what field? bro you gotta specify

GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE GEMINI IS BETTER THAN CLAUDE change my mind.

@hamoudi_yo I hope so, but the fact that they kept delaying made me concerned

@janedoe2ooo -it should be as good as 4.8 and 5.5 it's Delayed from June to end of July But it should be Good in real World Use not in Benchmarks

🚨Gemini 3.5 Pro Get delayed again 🤦To End of month Google not Needs to make The Best Model they Need To make a model as good as opus 4.8 and GPT 5.5 and Fix there Tools So People Start Using there Models And Tools and they Get training Data To make the Best Model in World

@knileshh Then why they keep delaying the release? It's definitely not because "oh it's too powerful we can't release it now we have to nerf it"

Gemini 3.5 pro

@TimTeaFan @synthwavedd Nah I don't wanna wait

@synthwavedd I think at this point it would be better to just skip 3.5 all together and release Gemini Pro 4 this fall.

In a move that I'm sure surprises absolutely nobody, Gemini 3.5 Pro has been delayed (again!), this time tentatively aiming for end-of-month 🤦‍♂️

@Hong60282445 @ajs6888 @JustinGorya I do believe so, but seeing core members leave one after one is also concerning

@ajs6888 @JustinGorya surely not, xAi just turned it around from nothing, and meta after just 2 releases is seemingly back. yes google has been out of the loop but eventually they will pull it together just a matter of when

its sad.. Gemini 3.5 pro is being delayed AGAIN. It feels like Deepmind really lost the game in this year. Gemini 3 was such a great model - but got so bad with Gemini 3.1 and even worse with 3.5.. I hope they will make a massive comeback with Gemini 4. Omni flash is still a nice video editing model. But at the same time, SpaceXAI and Meta made a huge comeback. Grok 4.5 and Muse Spark 1.1 are really great models. Anthropic, OpenAI, are still leading but the gap is closing fast.

@loglike_app I guess that means the model right now is pretty lame

GoogleがGemini 3.5 Proを来週17日に延期、アーキテクチャを丸ごと作り直すそうです。マイナー更新ではなく作り直しを選ぶあたり、各社の競争がかなり激しくなっているのを感じます。