Susid Sprava

@baseten @opencode This tweet was generated by AI `day; slow` 🤫

10M developers use @opencode every month. This means the experience has to feel the same every hour of every day; slow or inconsistent inference breaks productivity and flow. Enter Baseten. With Baseten's Model APIs, OpenCode achieves 5x faster TPS, sub-second TTFT, and 33% blended cost savings passed directly to users through cache token pricing.

@bridgemindai @aipulseda1ly after quantization

@aipulseda1ly There is no way I can run it, its 550b parameters

Nemotron 3 Ultra is live on OpenRouter and it's free right now. NVIDIA's first open frontier model built for agents. The specs: 550B params / 55B active MoE 1M context window 300+ tok/s It's open weights, which sounds like you can run it locally. You can't, realistically. 550B parameters needs serious datacenter GPUs, not a desktop. So the real story is the free hosted tier. Frontier class reasoning at $0 in and $0 out, while it lasts. For anyone building agents, that's the unlock. Test it before the free window closes. Dropping it into Hermes today. Will report back with real numbers.

@Prince_Canuma @GoogleDeepMind That's not possible. Gemma only displays text.

@sosichest @GoogleDeepMind Text + image + audio

🚀 Gemma 4 12B is here! We partnered with @GoogleDeepMind to bring and optimize their new dense and unifed multimodal model for Apple Silicon. ◈ 12B dense · 256K context ◈ Thinking mode (built-in reasoning) ◈ Vision: dynamic res, OCR, UI + charts ◈ Native audio: ASR + speech translation ◈ Function calling for agents ◈ Text + image + audio, interleaved Runs local. Get started now ⚡ > uv pip install -U mlx-vlm github.com/Blaizzy/mlx-vlm

Successfully wasted 5 years in Crypto.

Qwen 3.7 Max is now on Hyper. We waited a week to make sure it met our standards for performance and zero data retention. This one is impressive. Try it here: ⚡️ hyper.charm.land

Because humanity would try to put their d!ck in it. And if they did, they wouldn't be breathing anymore.

@CommandCodeAI Protonmail - it's temp email?

@sosichest We don’t allow email domains that are temp emails. Been getting a lot of abuse through that. Sign up with gmail?

People keep asking if this is a typo. It’s not. commandcode.ai/pricing

@r0ktech Before you use it, show your passport to the camera. This is absurd. Zuckerberg is losing his mind.

Who’s uses Meta AI ? 😹

@bridgemindai Be objective. Don't underestimate the models based on your own subjective opinion.

MiniMax M3 is free right now in OpenCode. On paper it beats GPT 5.5 on SWE-bench. But benchmarks get benchmaxed, so I trust them about as far as I can throw them. I am testing it live on real work right now. Real codebase, real TODOs, no cherry picking. We will see if the score holds up or falls apart. Report coming.

@Havira_aa @grok Not bad. Who is?

@TeksEdge @OpenRouter I suspect they implemented 4-bit quantization, and the inference became so cheap that there’s no point in charging people those few cents.

Holy💩! I didn’t get the memo. Did Kimi K3.0 release? Because Kimi K2.6 is free in @OpenRouter - what?

BTC 70K BTC 70K in 2021 in 2026

@kilocode What nonsense. Even the free version of Deepseek Flash can handle this problem.

We gave Grok Build 0.1 one prompt: build a webhook delivery service in TypeScript, Bun, and SQLite. It planned it, built it, and shipped a working demo. Total cost: $1.65. Zero tool-calling failures. Here's exactly what happened, file by file.

@bridgemindai Complaining that Claude has cut the limits and canceling the subscription... in 3... 2...

UltraCode is running over 100 Opus 4.8 agents right now. This is the future of software development. Not 5 agents. Not 10. Over 100 working simultaneously across the entire BridgeMind ecosystem. This is what vibe coding looks like at scale. One person orchestrating a swarm of AI agents building, debugging, and shipping in parallel. We are so early.

@jcarlos2001 Once everyone has jumped on, the team needs to walk a few more meters.

Si jugaste a ésto, ya eres del grupo de los veteranos ...

WTF?! This changes image generation forever! 🤯 PrismML just released Binary and Ternary Bonsai Image 4B! That's right, 1-bit diffusion models are here. Only ~3GB in size (FLUX.2 Klein 4B is 16GB). The most shocking part? It can run 100% locally in your browser. Try it now! 👇

@xenovacom That's impossible!!!

@thdxr ???

this is the best model i've ever used