|
Introducing Teel – Turbo-Efficient Execution Library for Agents (but not only)
|
Shahar Lev, System Architect |
2026-11-05 |
2,513 |
--
|
|
Post-training case study: Reaching frontier-level intelligence on global payroll with Deel
|
[email protected] |
2026-11-04 |
1,736 |
--
|
|
Post-training case study: Reaching frontier-level attack resilience with Alice
|
AI21 Labs |
2026-11-02 |
2,077 |
--
|
|
Why we built a benchmark before training for instruction following
|
Eran Goldstein, Algorithm Developer |
2026-10-25 |
1,724 |
--
|
|
How we cut GitLab Duo Code Review’s cost by 98%
|
Stephane Benayoun |
2026-10-22 |
2,306 |
--
|
|
Blog | AI21
|
AI21 Labs Team |
2026-10-15 |
263 |
--
|
|
The hidden mismatch: Evaluating coding agents on the work developers actually do
|
Eli Lepkifker, Algorithm Developer |
2026-10-14 |
1,876 |
--
|
|
The verifiability litmus test for agent design
|
Yuval Belfer, Sr. Developer Advocate |
2026-10-06 |
2,076 |
--
|
|
From manual negotiation to automated scheduling: How AI21 manages its GPU fleet with Kueue
|
Asaf Ben-Tovim, DevOps Engineer |
2026-10-04 |
2,825 |
3
|
|
You don’t need a frontier model. You need a verifier.
|
Guy Freund, Algorithm Developer |
2026-08-19 |
1,706 |
--
|
|
Better and cheaper together: Open models explore, frontier models patch
|
Eli Lepkifker, Algorithm Developer |
2026-07-15 |
1,164 |
--
|
|
Improving Best-of-N with Budget-Aware Execution for SWE Agents
|
Assaf Gerner, Algorithm Developer |
2026-07-07 |
2,054 |
--
|
|
Token spend isn’t going down. You need more than naive routing to manage it
|
Ori Goshen |
2026-06-25 |
727 |
--
|
|
Tipping the scales: Merging weak agents into a state-of-the-art deep researcher
|
AI21 Labs |
2026-06-24 |
1,464 |
--
|
|
First scale, then enrich: How the right execution strategy helped us reach state-of-the-art on SWE-rebench
|
Eli Lepkifker, Algorithm Developer |
2026-06-04 |
2,877 |
--
|
|
Reproducing Variance: Caching in Agentic LLM Pipelines
|
Guy Lichtenfeld, Software Engineer • Orchestration |
2026-05-13 |
2,024 |
--
|
|
Reaching SOTA Without Breaking the Bank: Using AI21 Maestro to optimize deep research agents
|
Amnon Morag |
2026-04-28 |
3,163 |
1
|
|
All that glitters: When “gold-like” answers mask functional failures on coding agent benchmarks
|
Oded Avraham, Algorithm Developer |
2026-04-14 |
3,464 |
--
|
|
Engineering the subconscious: Why Claude Code isn’t enough to build AI systems
|
Idan Rejwan |
2026-04-05 |
1,344 |
--
|
|
Stride and prejudice: How a 32-bit overflow corrupted a CUDA kernel (and stayed hidden for weeks)
|
Tamer Ghattas, DL Engineer |
2026-03-25 |
2,000 |
2
|
|
Mind the gap
|
Yuval Belfer, Senior Developer Advocate @ AI21 |
2026-03-17 |
1,939 |
--
|
|
Where enterprise AI deployments actually get stuck
|
Idan Rejwan, Director of Solutions |
2026-03-10 |
1,138 |
--
|
|
Modular intelligence: a human-like model for agent orchestration
|
Julie Fadlon, PhD, Head of Human Alignment & Standards |
2026-02-26 |
1,401 |
--
|
|
Reducing LLM training waste with model-agnostic padding minimization
|
Roee Hendel, Algorithm Developer |
2026-02-11 |
2,570 |
--
|
|
Go big or go OOM: the art of scaling vLLM
|
Ella Neiman, Engineering Team Lead |
2026-02-05 |
2,315 |
--
|
|
One token to corrupt them all: a vLLM debugging tale
|
Asaf Gardin, Senior Software Engineer |
2026-01-29 |
3,349 |
--
|
|
Chunk size is query-dependent: a simple multi-scale approach to RAG retrieval
|
Niv Granot, Algorithms Group Lead @ AI21 |
2026-01-29 |
1,700 |
5
|
|
Closing the parsing gap: reaching SOTA RTL parsing by leveraging LTR capabilities
|
Yuval Peleg Levy, Algorithm Engineer |
2026-01-22 |
2,412 |
--
|
|
When sleeping in saves you money: dynamic data snoozing for efficient online RL
|
Daniel Gissin, Tech Director, Post-Training Lead |
2026-01-22 |
2,247 |
--
|
|
Boring isn’t easy
|
Tal Guttman, Head of Product Inspiration |
2026-01-15 |
846 |
--
|
|
Introducing Jamba2: The open source model family for enterprise reliability and efficiency
|
Joanna Kramer |
2026-01-08 |
887 |
--
|
|
How to scale agentic evaluation: Lessons from 200,000 SWE-bench runs
|
Yaron Sternbach, VP Engineering |
2026-01-08 |
1,384 |
4
|
|
Shared state, no drama: Scaling state-modifying agents with MCP workspaces
|
Eran Gat, System Lead |
2026-01-07 |
1,230 |
--
|
|
Elevating long-horizon agentic tasks with orchestrated Test-Time Compute
|
Or Dagan, Chief Product & Strategy Officer |
2026-01-07 |
1,974 |
--
|
|
Enterprise AI after the hype curve
|
Yehoshua (Shuki) Cohen, VP Data, AI Evangelist |
2025-12-31 |
1,017 |
--
|
|
Expanding enterprise access to open-source AI: Our partnership with Together AI
|
AI21 |
2025-11-26 |
514 |
--
|
|
AI21 Labs recognized as an Emerging Visionary in two Gartner® Emerging Market Quadrants for Generative AI
|
AI21 |
2025-11-18 |
680 |
--
|
|
AI21 Maestro’s accuracy fix for RAG’s blind spots
|
Amnon Morag, VP Product & Market Strategy @ AI21 |
2025-11-12 |
1,447 |
--
|
|
The latency test: Jamba 3B vs Qwen3 4B 2507
|
AI21 |
2025-11-11 |
123 |
--
|
|
Build a reliable AI agent in days with AI21 Maestro
|
AI21 |
2025-11-09 |
598 |
--
|
|
Five questions in five Minutes: Dr. Avi Tsur on AI in healthcare
|
AI21 |
2025-10-22 |
589 |
--
|
|
Introducing Jamba Reasoning 3B: Tiny Model, Huge Possibilities
|
AI21 Editorial Team |
2025-10-08 |
866 |
--
|
|
AI in finance: Progress, barriers, and the road ahead
|
AI21 |
2025-09-02 |
831 |
--
|
|
Knowledge Agents for the Enterprise: From Retrieval to Real Work
|
AI21 |
2025-08-25 |
244 |
--
|
|
Attention was never enough: Tracing the rise of hybrid LLMs
|
Dor Schwartz |
2025-08-05 |
1,190 |
2
|
|
“AI all the work”: Inside Ori Goshen’s RAISE Summit keynote
|
AI21 |
2025-07-30 |
229 |
--
|
|
Why Grounding is (Still) the Bedrock of Enterprise AI
|
AI21 Editorial Team |
2025-07-30 |
1,883 |
--
|
|
Karpathy’s Leash
|
AI21 |
2025-07-23 |
858 |
--
|
|
Automating Product Descriptions in Real Time: Clear the Content Debt, Win the Cart
|
AI21 |
2025-07-22 |
1,041 |
--
|
|
Monitoring Compliance in Real Time: Reduce Risk, React Faster
|
AI21 |
2025-07-21 |
1,173 |
--
|
