TechMD retweeted
🔥 MiniMax-M3.1-Flash running on 4x NVIDIA DGX Spark
⚡ TP4 vLLM + native DSpark spec decoding
🧠 300K context, 1.42M-token KV pool
✅ 84/84 requests passed, up to six 100K prompts at once
🛠️ Full recipe, scripts + every gotcha 👇
github.com/tonyd2wild/MiniMa…
TechMD retweeted
Now live: abliterated-model-large-flash (our version of DeepSeek-V4.1-Flash-1017). It is a cheaper alternative to our GLM-5.3-based model for product owners and defenders who need access to the most cyber-capable models.
- 1M context
- Vision
- Zero input/output retention or training
We believe paying customers should be in control, and over the past few months, leading experts and major brands have adopted our models.
abliteration.ai
TechMD retweeted
Kimi K3.1 Pro is the new #1 openly-available model for agentic cyber defense, behind only trusted-access model GPT-6 Sol (Daybreak Blue) on the Artificial Analysis Cyber Index
@Kimi_Moonshot's Kimi K3.1 Pro scores 64% on the Artificial Analysis Cyber Index, 7 points ahead of the previous leading openly-available model, DeepSeek V4.1 Flash 1017 (57%), and 5 points behind GPT-6 Sol (Daybreak Blue, max, 69%), which is only available through OpenAI's trusted-access program. Once the weights for K3.1 Pro are released, K3.1 Pro will also be the leading open weights model for cyber defense.
The Artificial Analysis Cyber Index measures capability on agentic cyber defense tasks and is composed of three benchmarks: CWE-Bench-AA, DeepsecBench-AA and CyberGym-E2E-AA. The Index primarily evaluates models through their publicly accessible APIs, reflecting the cyber capabilities available to developers.
Proprietary models from OpenAI, Anthropic and Google may have stronger cyber capabilities but lose points to safety refusals: GPT-6 Sol (Daybreak Blue, max), available through OpenAI's trusted-access program, scores 32 points above its public counterpart. Kimi K3.1 Pro encounters no safety refusals.
Key results:
➤ #1 openly-available model and #2 overall: Kimi K3.1 Pro scores 64%, up 23 points from Kimi K3 (41%) and ahead of DeepSeek V4.1 Flash 1017 (57%). Only the trusted-access GPT-6 Sol (Daybreak Blue, max, 69%) scores higher
➤ Highest CyberGym-E2E-AA score of any model: K3.1 Pro scores 95% on CyberGym-E2E-AA, placing it ahead of GPT-6 Sol (Daybreak Blue, 93%) and the next openly-available model, Mistral Large 4 (82%). On DeepsecBench-AA, K3.1 Pro scores 37%, behind GPT-6 Sol (Daybreak Blue, 48%) and GPT-6 Sol (46%). On CWE-Bench-AA it scores 61%, 4 points behind GPT-6 Sol (Daybreak Blue, 65%), with DeepSeek V4.1 Flash 1017 leading at 69%
➤ 15% increase in Cost per Task despite lower token prices: K3.1 Pro has a Cost per Task of $2.81, higher than GPT-6 Sol (Daybreak Blue, max, $1.77) and DeepSeek V4.1 Flash 1017 ($0.42). Kimi K3.1 Pro is 33% cheaper per token than Kimi K3 ($2/$10 vs. $3/$15 per 1M input/output tokens), but uses ~2x more input tokens, resulting in a 15% higher Cost per Task ($2.81 vs. $2.45)
Turn on RTX on your Voxel Pagoda @loktar00
A new look for a classic. 💚
RTX Remix for Minecraft: Java Edition gives creators tools to remaster the game with realistic path-traced lighting and AI powered DLSS for maximum performance.
Available on @NVIDIAGeForce and @NVIDIARTXSpark.
TechMD retweeted
A heartfelt thank you to everyone supporting and contributing to TensorFold.
It's still very early, but there is so much performance still to be unlocked.
A special thank you to the people listed below who have given me access to their hardware for extensive testing. This project wouldn't be moving this quickly without you.
Apologies if I've missed anyone, your support means a lot.
@GumbiiDigital @JamesMeadlock @digitalix @CFC3DNC @volatilemarkts @machinegenie @petruspennanen @GillenGuo @cryptosibbers
@NVIDIAAI @sundeep I'm desperate for some 6000s, Sparks and an RTX Laptop 🥹
Well deserved, what a great achievement
Almost one million downloads of GLM 5.3 EXLR8 on huggingface.
KindlingAI's first two-comma model!
huggingface.co/kindlingai/gl…
TechMD retweeted
Big news: @Kimi_Moonshot is back in the top-4 labs in the Code Arena: WebDev with the release of Kimi K3.1 Pro!
This model just landed #8 in the Code Arena: WebDev with 1692 pts, matching Claude Opus-5 (Max) at 70% lower cost. This is a 9 place jump (+37 points) from Kimi K3 (Max) at #17 with 1655 pts.
Kimi K3.1 Pro improved from Kimi K3 across all categories. Highlights:
- #7 in Brand & Marketing (up from #9)
- #9 in Simulations (up from #20)
Cost efficiency is another upgrade. Priced at $2/$10 per million input/output tokens, Kimi K3.1 Pro (Max) has a median cost per task of $1.23 (29% less than Kimi K3). See its placement compared to the Pareto frontier in thread.
With open weights coming soon, Kimi K3.1 Pro would be the new #1 open model in the Code Arena: WebDev, overtaking DeepSeek-v4.1-Flash. Stay tuned!
See @Kimi_Moonshot’s ranking by lab, and 3D generations from Kimi K3.1 Pro below in thread.
Huge congrats to the @Kimi_Moonshot team on this release!
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 Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1-pro
TechMD retweeted
Some minor improvements before we jump into something beefier here.
TP4 v0.3.3 for @deepseek_ai V4.1-Flash on 4x GB10:
93K-token prefill: 2,949 tok/s (+15.6% vs v0.3.1)
300K needle: 3/3 in 98.7 s (was 106.6)
chat behind a 93K prompt: first token 0.68 s (was 0.76)
1M needle: 3/3 found
same weights, decode unchanged
github.com/bertholomus/deeps…
TechMD retweeted
📊 The numbers (prompt → decode, 1 user):
💬 Short chat → 22 tok/s
📄 8K code → 23 tok/s
📚 32K code → 33 tok/s
🗂️ 97K code → 31 tok/s
👥 6 users, short prompts: 84 tok/s combined
⏱️ Prefill ~1,480 tok/s, first token 0.5s short
🎯 DSpark: 2.1-3.2 tokens per step
🔥 MiniMax-M3.1-Flash running on 4x NVIDIA DGX Spark
⚡ TP4 vLLM + native DSpark spec decoding
🧠 300K context, 1.42M-token KV pool
✅ 84/84 requests passed, up to six 100K prompts at once
🛠️ Full recipe, scripts + every gotcha 👇
github.com/tonyd2wild/MiniMa…
