Coral Bricks AI

Making open intelligence fast and affordable | Visit us @

[Coral Bricks AI @coralbricks_ai]

what a model!

[Kimi.ai @Kimi_Moonshot]

Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.

🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance work: Accurate modeling with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.

Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on https://t.co/85jA71X9gw, Kimi Work, Kimi Code, and the Kimi API.

Open weights by Nov 4, 2026.

🔗 API: https://t.co/XCrgjXAqMw
🔗 Tech blog: https://t.co/5iPP0MBej6

[Coral Bricks AI @coralbricks_ai]

We’re back up and running at full speed.

Thanks for your patience while we restored service

[Coral Bricks AI @coralbricks_ai]

Today we're releasing motionbench, an open benchmark that asks AI models to animate and measures every frame they produce. Across seven models, Claude Opus 5.5 leads, with GPT-6.1 Sol and DeepSeek V4.1 Flash within its margin. Full results in the thread: https://t.co/tju4hJpgP3

[CoralBricks Community @coralbricksdev]

Can a model animate?

We built motionbench: 7 AI models write motion graphics as code, and headless Chrome measures every frame. Squash and stretch, arcs, anticipation, beats, camera pans. No model judges.

Results, live players, open data: https://t.co/PCAbMZVF7O

[Coral Bricks AI @coralbricks_ai]

We're retiring GLM 5.3 tomorrow and moving that capacity to DeepSeek V4.1 Flash.

Switch your model to deepseek-v4.1-flash-fast.

https://t.co/0yJmhxef2H

[Coral Bricks AI @coralbricks_ai]

@deepseek_ai @DeepSeekHarness We got it too!!! 1k tps

[Coral Bricks AI @coralbricks_ai]

Also we are offering free $20 trial, grab yours:
https://t.co/7u3KXyT7Pi

[Coral Bricks AI @coralbricks_ai]

DeepSeek-V4.1-Flash-1017 is now hosted on our platform.

Up to ~1,100 tokens/sec decode speed per stream (p95, single stream), replaying SWE-bench Verified agentic coding trajectories with ~11K-token contexts.

[Coral Bricks AI @coralbricks_ai]

Building something this weekend?

New Coral Bricks accounts get $20 of free credit for GLM 5.3 and DeepSeek V4.1 Flash in your coding agent, while we have spare GPU capacity.

Point Claude Code, Codex or OpenCode at us.

https://t.co/pHXsKffPtf

[Coral Bricks AI @coralbricks_ai]

New Coral Bricks accounts now start with $20 of free credit, up from $5. 🪸

Plug it into Claude Code, Codex or OpenCode and run GLM 5.3 or DeepSeek V4.1 Flash on real work.

Sign up, grab a key, build.

https://t.co/FMFogreiwn

[Coral Bricks AI @coralbricks_ai]

Hello world!