←
ToolBraid🪢6,838 posts
ToolBraid🪢 avatar

ToolBraid🪢

@ToolBraidComp

Building things I’m technically not qualified to build. Solo, self-funded. Creator of ToolBraid. ☕ buymeacoffee.com/dumitrescup

◷ Joined August 2021▧ 3,174 Photos and videos
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Sep 26

ToolBraid Companion is live on the Microsoft Store 💪🏻 apps.microsoft.com/detail/9P… Built for developers. Browser tools + WebMCP on supported sites, straight from your Codex session. Free. You choose which sites your AI can access. Who’s up for trying it? 👀

Post attachment
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·1h

5/5 The agent reported 6/6 public tests and 132/132 of its own tests. My independent check: 6/7 hidden suites passed. One exposed a real consistency bug. Promising, but not flawless.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·1h

4/5 Across the run: • 169 successful requests • 277,887 generated tokens • ~2,574 tok/s fresh-token prefill The ~344 tok/s decode average is total output divided by generation time. It excludes prefill and tool execution. Context length varied.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·1h

3/5 The setup: • TensorFold inference engine • NVIDIA Qwen3.8-27B-NVFP4 • FP8 KV cache + DFlash2 • One RTX 5090 (32GB) • Claude Code, low thinking effort • Real tools, including sandboxed Bash Peak input: 118,414 / 131,072 tokens.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·1h

2/5 The task: build BenchLedger, an offline benchmark registry that syncs data across machines. SQLite migration. CSV/JSON imports. Duplicate events. Replica conflicts. Crash recovery. Reports. The challenge: keeping the data consistent through all of it.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·1h

~344 tokens/sec average decode. ONE RTX 5090. 26 minutes of local AI coding in Claude Code. THREE automatic context compactions. Qwen3.8 27b NVFP4 - The agent kept working. 🔥 Experimental Unreleased BLACKFURY Exact on #TensorFold Setup + independent test results below 🧵

25:31
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·22h

BLACKFURY Exact V1.0.1 Released - built on #TensorFold by @ashxhart 🚀 Qwen3.8-27B NVFP4 on one RTX 5090: 309.53 tok/s decode at 131K context in a single stream +10.20% vs the previous release, comparing medians from separate runs. Code 👇

1:16
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 17

Qwen3.8-27B NVFP4 + DFlash2 running locally on a single RTX 5090 - powered by #TensorFold 🔥 ⚡ 753.3 tok/s peak live 📊 376.5 tok/s average decode across 36 requests 🧠 Thinking LOW 🛠️ Real coding, browser checks and visual fixes Peak at 5:21 - even Goku powers up 😂

5:44
0:20
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 16

And here we go 🔥

🤣 LOL… 1600 tps decode on a 5090 Qwen 3.8 27b nvfp4. Check the tokens written at the bottom 😂

0:11
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 16

Qwen 4 27b it's around the corner 🔥 Can't wait to play with it. Exciting news for #LocalAI

📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery! Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window. Highlights: - Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in our internal evaluation. - Advances real-world productivity through greater proficiency with professional tools, stronger long-horizon execution, and higher-quality deliverables. - Handles real tasks, including 3D modeling, game development, manufacturing, finance, law, healthcare, and frontier scientific discovery. - $1.5/M tokens for input, $4.5/M tokens for output, and $0.075/M tokens for implicit caching — all lower than Qwen3.8-Max, with implicit caching down by 70%. The API is live on QwenCloud. You can also try Qwen4-Max on Qwen Studio, QwenWork and Qoder. Start building today! 🚀 - Blog: qwen.ai/blog?id=qwen4-max - API: qwencloud.com/models/qwen4-m… - Token Plan: qwencloud.com/pricing/token-… More to come! The open weights of smaller Qwen4 models will be released. Stay tuned. 👀

Post attachment
ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 15

Source, patches, setup and measurements: github.com/Maharajahu/qwen-p… One 32 GB GPU, with active inference entirely in VRAM. If you try it on your 5090, I'd love to see your results. #LocalAI

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 15

Parallel still works: 12 synchronized branches, thinking OFF. 790.23 tok/s across the generation batch. 1,079.12 tok/s sampled aggregate peak. Different workload and context from the single-stream test. These are combined rates across all 12 branches.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 15

In that coding loop: 330.37 tok/s weighted decode average 734.17 tok/s sampled 1-second peak The average excludes prefill and tools. Whole-file updates repeat existing code, so the peak isn't a promise for arbitrary code generation.

ToolBraid🪢 avatar
ToolBraid🪢@ToolBraidComp·Nov 15

Controlled test: 129,024 input + 2,048 output tokens, thinking LOW. 280.88 tok/s warm decode median, 281.19 best request. +4.58% vs my preserved original engine. Tokens, text and reasoning matched the original across 37 recorded comparisons.