Snorkel contributed: • An agentic RL eval environment • FinQA-Reasoning dataset • Finance Reasoning benchmark Full technical breakdown from the rLLM team + our enterprise takeaways here:
LLMS
-
Tool Discipline Improves AI Reasoning Better Than Scale
By
–
Key insight: tool discipline > scale. Instead of complex multi-table training, we reinforced reliable tool use on simple queries — and saw transfer to 12-step reasoning tasks (59.7% Pass@1).
-
4B Model Outperforms 235B on Financial Reasoning Tasks
By
–
A 4B model > 235B on financial reasoning. We partnered with @rllm_project to fine-tune Qwen3-4B-Instruct-2507 — and it outperformed Qwen3-235B-A22B on expert-curated financial benchmarks.
-
V3 leads in cost with rumors of wildly cheap tokens
By
–
V3 lead the pack in some respects when it came out, certainly competitive in cost. Rumors are pointing to wildly cheap token costs, TBD obviously. More of a 'what-if' scenario.
-
From 10% to 76%: LLM success rate in data analytics
By
–
I'm sure comments will be around hallucinations and how its hard to use LLM's in data analytics. This is a branded chart, BUT look at the trajectory here from 10% success rate in sept 2025 to 76% in February 2026 In AI land, you need to really shift your perspective to keep up
-
DeepSeek V4 could slash frontier model costs dramatically
By
–
DeepSeek V4 is dropping soon, and it could be a full reset of the leaderboards. What happens to the API revenue of Anthropic/OpenAI when an open source model slashes the cost of 1M frontier model tokens from $15 to $0.27?
-
Train LLMs in VS Code for Free with Unsloth and Colab
By
–
Now you can train LLMs in VS Code for free!
— Akshay 🚀 (@akshay_pachaar) 18 février 2026
This guide by Unsloth shows you how to connect any fine-tuning notebook in VS Code to a Colab runtime.
Train locally or on a free Google Colab GPU.
Guide: https://t.co/ChBeoyw6Mh
GitHub: https://t.co/smDHEA1tI8 pic.twitter.com/ma2qgblO8eNow you can train LLMs in VS Code for free! This guide by Unsloth shows you how to connect any fine-tuning notebook in VS Code to a Colab runtime. Train locally or on a free Google Colab GPU. Guide: https://
unsloth.ai/docs/get-start
ed/install/vs-code
…
GitHub: https://
github.com/unslothai/unsl
oth
… -
GLM-5 Open Weights Model Beats Gemini-3-Pro
By
–
GLM-5: A new SoTA open weights, that BEATS Gemini-3-Pro!? A big push from vibe coding -> agentic engineering, with it being able to plan, act, and iterate over long workflows, not just spit out code. GLM-5 scales to 744B params (40B active MoE) and targets 200k contexts, with
-
Claude Sonnet 4.6 Enables Advanced Research Paper Analysis
By
–
Introducing Claude Sonnet 4.6 for understanding research papers 🚀
— alphaXiv (@askalphaxiv) 18 février 2026
Highlight any section of a paper to ask questions and “@” other papers for quick context, comparisons, and benchmark references pic.twitter.com/60eq7dLtEzIntroducing Claude Sonnet 4.6 for understanding research papers Highlight any section of a paper to ask questions and “@” other papers for quick context, comparisons, and benchmark references
-
Gemini 3 and GPT 5 Now Available on alphaXiv
By
–
Now available in addition to Gemini 3 and GPT 5. Check out http://
alphaXiv.org! Alternatively check out our Chrome extension