Really impressive for a 15B model, basically a self-hosted Gemini flash.
ServiceNow has released Apriel-v1.5-15B-Thinker, a 15B open weights reasoning model that leads our Small Models category (<40B parameters)
💼 Overview: Apriel-v1.5-15B-Thinker is a dense, 15B parameter open weights reasoning model. This is not the first model ServiceNow has released but is a substantial jump in intelligence achieved compared to past releases
🧠 Intelligence: The model scores 52 in the Artificial Analysis Intelligence Index. This puts it on par with DeepSeek R1 0528, which has a much larger 685B parameter architecture. ServiceNow’s model scores particularly well within important behaviors for enterprise agents, such as instruction following (62% in IFBench, ahead of gpt-oss-20B, reasoning) and multi-turn conversions & tool use (68% in 𝜏²-Bench Telecom, ahead of gpt-oss-120B, reasoning). This makes it particularly well-suited to agentic use cases, which was likely a focus given ServiceNow is active in the enterprise agents space
⚙️ Output tokens and verbosity: The model produces a large number of output tokens even among reasoning models - using ~110M combined reasoning and answer tokens to complete the Artificial Analysis Intelligence Index
🖥️ Access: No serverless inference providers are yet serving the model, but it is available now on Hugging Face for local inference or self-deployment. The model has been released under an MIT license, supporting unrestricted commercial use
ℹ️ Context window: The model has a native context window of 128k tokens.
Congratulations to @ServiceNowRSRCH on this impressive result!
Ah, if it's uppercase letters (A-Z) and numbers (0-9) only—36 characters—then 36^6 ≈ 2.18 billion possibilities.
Brute forcing online: At 1 attempt/sec, average ~34.5 years. At 100/sec (if allowed), ~4 months. Still, rate limits would block it. Best guess: 'GROKAI'.
Masterpiece about context engineering
New on the Anthropic Engineering Blog: Most developers have heard of prompt engineering. But to get the most out of AI agents, you need context engineering.
We explain how it works: anthropic.com/engineering/ef…
Silent Sense: This is a real breakthrough that can change our world and how humans and machines communicate.
Nacer.bs retweeted
I reverse engineered @LucidEditor’s new "AI-native word processor" - and their client-side code exposes their entire system prompt.
What I found is absolutely unhinged 🧵
Hey @elder_plinius, I'm sure you will love this one: piped.video/vEioDeOiqEs?si=zriM…
Hey @ArtificialAnlys, could you rerun the openai o3 benchmarks and check for metric updates?
Lot of people suspecting that the model was quatized.
