Laguna XS 2.1
PoolsideCompact
ThinkingTool Use
About this model
Laguna XS 2.1 is the latest coding agent model in the 33B-A3B category from [Poolside](https://poolside.ai/) and a step forward from their Laguna XS.2 model (released in April 2026). It combines...
Performance Tier
Compact
Laguna XS 2.1 is a compact model from Poolside : optimized for speed and affordability.
Small, fast, and affordable. Optimized for speed and low cost, great for high-volume or simple tasks.
Pricing
This model is included in Elosia plansEco
Minimal cost. Ideal for very high volume or simple tasks.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.060 |
| Output (completion) | $0.120 |
| Cache read | $0.030 |
Capabilities
Context Length262K
Max Output Tokens33K
TokenizerOther
Inputtext
Outputtext
Release DateJuly 2, 2026
Benchmarks
General Intelligence
MMLU
Not reported
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
SWE-bench Verified
70.9%
SWE-bench Multilingual
63.1%
Agentic
SWE-bench Pro
47.6%
Terminal-Bench 2.0
37.5%
Where does Laguna XS 2.1 stand?
Compare its performance index against every other model.
Recommended Use Cases
Coding
Strengths
- Compact Mixture-of-Experts (33B total / 3B active) for local agentic coding
- Strong code performance for its size (SWE-bench Verified 70.9%, SWE-bench Pro 47.6%)
- Improved multilingual code (SWE-bench Multilingual 63.1%, up 5.4 points over XS.2)
- Open weights under Apache 2.0, runs on a local machine, 256K-token context
Limitations
- Code specialist: no published general-knowledge or math benchmarks
- Benchmarks are self-reported by Poolside (Harbor framework)
Frequently asked questions
Resources
This model may use your data for training
