thinkthinking

Enjoying Vibe-Working 💫
Co-Founder & Product Lead @ZenMuxAI | AgentOS /
AI Product Manager @AntGroup | @Baidu_Inc /
Vision Perception Engineer @CATL_Official

[thinkthinking @thinkthinking_]

有一说一，Claude Opus 4.8 写提示词的能力是真强！同样的生图 Skill ，同样用的 GPT-Image-2 模型。
Claude Opus 4.8 写出来的提示词生成的图片效果质量要比 GPT 5.5 好一截！画面更高级，更有意境！

[thinkthinking @thinkthinking_]

🚀 ZenMux now supports Claude Opus 4.8 @AnthropicAI
Free tier users can try it out at no cost. Come experience the latest model! 🎉

[Claude @claudeai]

Introducing Claude Opus 4.8: it builds on Opus 4.7 with sharper judgment, more honesty about its own progress, and the ability to work independently for longer than its predecessors.

Available today at the same price.

[thinkthinking @thinkthinking_]

RT @ZenMuxAI: Qwen 3.7 Max is now live on ZenMuxAI.
Available via both:
- OpenAI-compatible API
- Anthropic-compatible API
Built for the ag…

[thinkthinking @thinkthinking_]

@ZenMuxAI now supports the @OpenAI Image protocol.
You can call OpenAI image models natively through the OpenAI SDK:

Base URL:
https://t.co/0EgfB4kdXa

Supported models:
- openai/gpt-image-2
- openai/gpt-image-1.5

Docs:
https://t.co/bxYoZFKwtw

For agent workflows, install the ZenMux image generation skill:

npx skills add ZenMux/skills

[ZenMux @ZenMuxAI]

GPT‑5.5 Instant is now live on ZenMuxAI.

Available via both:
- OpenAI-compatible API
- Anthropic-compatible API

This is a surprisingly meaningful upgrade over GPT‑5.3 Instant:
- 52.5% fewer hallucinated claims
- much better self-correction
- shorter + denser responses
- stronger personalization & memory usage
- major gains in math / science / multimodal evals

One subtle but important change:
GPT‑5.5 Instant is much better at realizing its own reasoning mistakes and continuing to debug instead of stopping early.

That matters a lot for:
Coding, Agents, research, long-context workflows.

Now available on ZenMuxAI.

[thinkthinking @thinkthinking_]

DeepSeek-V4 is HERE FREE on @ZenMuxAI ! 🚀
API & Chat are free for a limited time – jump right in!
Two editions, one mission:
🔷 V4-Pro – max performance, built for agentic coding. Real‑world feel: better than Sonnet 4.5, close to Opus 4.6 (non‑thinking).
⚡ V4-Flash – smaller, faster, lower cost. Simple agent tasks? On par with Pro.
🔥 One API, all protocols – OpenAI Chat/Responses + Anthropic Messages. Use it inside Codex, Claude Code, OpenCode, etc.
PAYGO & Builder Plan already supported.
Try for free now → ZenMuxAI

#DeepSeek #DeepSeekV4 #AI #LLM #VibeCoding #OpenSource #ZenMuxAI

[DeepSeek @deepseek_ai]

🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.
🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at https://t.co/GCdiMzk1Dl via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: https://t.co/drlDrxkYtp
🤗 Open Weights: https://t.co/T13Y8i7SDM

1/n

[thinkthinking @thinkthinking_]

Gemini 3.1 Limited FREE just in ZenMuxAI

[ZenMux @ZenMuxAI]

Gemini 3.1 Pro launches on ZenMux | FREE limited trial
Now live on ZenMuxAI, compatible with OpenAI/Anthropic protocols.
Enhanced Pro model: reasoning, coding, multimodal, Agent.
API access available. Free tier enabled. See Figure 2.
Go!

[ZenMux @ZenMuxAI]

🚀 @Zai_org GLM-4.7 is live on ZenMuxAI
Use GLM-4.7 in Claude Code / Codex via ZenMux.
No code changes.
Benchmarks:
• SWE-bench Verified 73.8% (open-source SOTA)
• LiveCodeBench v6 84.9%
• HLE 42.8%
👉 https://t.co/DDwZzvHjUi

[ZenMux @ZenMuxAI]

🚀 @minimax_ai
MiniMax M2.1 is live on ZenMuxAI
Use M2.1 in Claude Code / Codex via ZenMux.
• Stronger multi-language coding (Rust/Java/Go/C++/TS…)
• Better agent instruction following & efficiency
• VIBE full-stack benchmark: 88.6
👉 https://t.co/nsIkFZ5XHU

[泊舟 @bozhou_ai]

看到新发布的AI增速榜单，ZenMux排名第12，增长525%。

在OpenRouter和大厂牢牢占据的市场里，ZenMux还能这么快速增长，我觉得不是运气。

最直观的原因很简单，模型新、稳定、延迟低，用户体验好。
具体来说：

模型上新超级快。之前NanoBanana Pro和现在Gemini 3 flash官方还不能用，他们平台已经可以接入了。

而且原生支持OpenAI、Anthropic、Vertex API，这意味着大家可以直接集成到Claude Code和Codex里用，无缝迁移。

后台体验做得很用心。可以清晰看到logs、usage、cost等维度，每分钱花到哪里了一目了然。这对任何一个用户来说都是刚需。

稳定性和延迟上没得说，这是基本盘。
然后ZenMux还做了保险机制。模型出现质量问题会进行补偿，这给了用户额外的安全感。长期来看，这能帮挖掘业务场景里的bad case，沉淀数据资产，形成产品迭代的数据飞轮。

我自己在用，最近也给公司推荐接入，特别是为不方便翻墙的同事体验顶级大模型。

产品基本盘稳 + 开发者友好 + 保险机制的长期价值+日常充值赠送20%

未来可期

[ZenMux @ZenMuxAI]

Doubao-Seed-1.8 has landed on @ZenMuxAI . ⚡️

Featuring stronger Multimodal Understanding and enhanced Agent Capabilities.

This model is engineered to deliver outstanding performance in complex, real-world tasks.

Experience the upgrade immediately: https://t.co/8TBqqRQpmj

[thinkthinking @thinkthinking_]

@aisuskii @ZenMuxAI You can use Gemini 3 Flash in https://t.co/cja6J5Owzb now！

[thinkthinking @thinkthinking_]

⚡️⚡️⚡️
Gemini 3 Flash Preview is now FREE to try on @ZenMuxAI — one of the first platforms to integrate it (ahead of official release).
Supports OpenAI / Anthropic / Google Vertex APIs.
Use it instantly in Claude Code & Codex.

[ZenMux @ZenMuxAI]

⚡️ Early Access: Gemini 3 Flash Preview.

We’ve integrated it ahead of the official rollout—and it’s currently FREE to try on @ZenMuxAI.

🛠️ Seamless Integration:Supports OpenAI / Anthropic / Vertex API natively. This means you can drop it directly into Claude Code & Codex today.

Be the first to test for free:👉 https://t.co/hduTxQzaXl

[thinkthinking @thinkthinking_]

@XiaomiMiMo @luo_fuli14427 latest model MiMo-V2-Flash is now live on @ZenMuxAI.
ZenMuxAI supports OpenAI & Anthropic–compatible APIs, drop-in usable in Claude Code / Codex — no code changes, just switch the endpoint.

[Xiaomi MiMo @XiaomiMiMo]

⚡ Faster than Fast. Designed for Agentic AI.
Introducing Xiaomi MiMo-V2-Flash — our new open-source MoE model: 309B total params, 15B active.

Blazing speed meets frontier performance.

🔥 Highlights:
🏗️ Hybrid Attention: 5:1 interleaved 128-window SWA + Global | 256K context
📈 Performance:
⚔️ Matches DeepSeek-V3.2 on general benchmarks — at a fraction of the latency
🏆 SWE-Bench Verified: 73.4% | SWE-Bench Multilingual: 71.7% — new SOTA for open-source models
🚀 Speed: 150 output tokens/s with Day-0 support from @lmsysorg🤝

🤗 Model: https://t.co/4Etm0yZKTL
📝 Blog Post: https://t.co/5zxmcDuB6o
📄 Technical Report: https://t.co/crac1YTLYl
🎨 AI Studio: https://t.co/nSReUs6QgW

[thinkthinking @thinkthinking_]

@XiaomiMiMo @luo_fuli14427 latest model MiMo-V2-Flash is now live on ZenMuxAI.

ZenMuxAI supports OpenAI & Anthropic–compatible APIs,
drop-in usable in Claude Code / Codex — no code changes, just switch the endpoint.
Key facts:

⚡ MiMo-V2-Flash — open-source MoE model for Agentic AI
• 309B total params, 15B active per token
• Frontier-level performance at significantly lower latency

🏗️ Hybrid Attention
• 5:1 interleaved Sliding Window (128) + Global Attention
• 256K context
📈 Performance
• Matches DeepSeek-V3.2 on general benchmarks
• Lower latency
🏆 SWE-Bench Verified: 73.4%
🏆 SWE-Bench Multilingual: 71.7% (open-source SOTA)

🚀 Speed: ~150 tokens/s

[thinkthinking @thinkthinking_]

⚡️⚡️⚡️

[ZenMux @ZenMuxAI]

The "Weekend Early Access" is over. 🏁

Black Friday Week is OFFICIALLY LIVE on @ZenMuxAI . 🏴

If you burned through your free Gemini 3 Pro quotas over the weekend, it's time to get serious.

Get 25% Bonus Credits on ALL recharges.
This is the best time to stock up for Q4 production workloads.

Scale up now: 👉 https://t.co/9u3dDzNHR8

P.S. This 25% bonus essentially gives you tens of millions of extra tokens for free.🎁