Press & links
Exabase Achieves Highest Reported Score on Leading AI Memory Benchmark Using a Smaller, Cheaper Model
Exabase Memory Engine Outperforms Rivals on Key AI Benchmark While Cutting Model Costs
Small Model, Big Results: Exabase Claims Top Score on AI Memory Benchmark
New AI Memory System Beats Larger Competitors Using Fraction of Computing Power
London Startup Achieves Best-in-Class AI Memory Results Without Expensive Models
Exabase M-1 Tops Leading AI Memory Benchmark With a Cheaper, Faster Model
Other links
LinkedIn: Introducing our first memory system: Exabase M-1.
Reddit: Introducing Exabase M-1: State-of-the-art AI memory with a smaller, cheaper model
Reddit: Hitting #1 on the leading memory benchmark (LongMemEval) with a smaller model (Gemini Flash)
Medium: Why Advanced AI Teams Are Moving Beyond Vector Databases to Exabase
We accidentally set a new frontier in AI memory. 96.4% using a smaller, cheaper model.
Exabase Sets AI Memory Benchmark With Efficient Model
Everything you need to give your agent context.
Get started in minutes.
Start for free
Read the docs
Everything you need to give your agent context.
Get started in minutes.
Start for free
Read the docs
Research
Product
Blog
Pricing
Docs
Examples
Log in
Start for free
Log in
Get started
