LLMpoolside

Poolside: Laguna XS 2.1

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...

Anyone in the Space can @-mention Poolside: Laguna XS 2.1 with the team's shared context - pooled credits, one chat, one memory.

All models

Starter is free forever - 1 Space, 100 credits/month, 1 MCP. No card.

Verdict

Laguna XS 2.1 is Poolside's compact code model optimized for fast, cost-effective completion tasks. With 262K context and $0.06/$0.12 per Mtok pricing, it undercuts larger code models by 10-20x while maintaining solid performance on autocomplete and small refactors. The trade-off is weaker reasoning on complex architecture decisions or multi-file refactoring compared to Sonnet or o1. Reach for this when you need responsive, budget-friendly code assistance across large codebases without the latency or cost of frontier models.

Best for

  • Real-time code autocomplete in editors
  • Cost-sensitive batch code analysis
  • Large codebase context retrieval
  • Inline documentation generation
  • Quick syntax fixes and formatting

Strengths

The 262K context window handles entire modules or multi-file diffs without truncation, making it practical for codebase-wide search and localized edits. At $0.06 input, you can feed thousands of lines without cost anxiety. The XS designation suggests inference speed optimized for interactive use — expect sub-second completions for typical autocomplete scenarios. Pricing is roughly 15x cheaper than GPT-4o and 8x cheaper than Claude Sonnet 4, making it viable for high-volume developer tooling.

Trade-offs

Without public benchmarks, performance on HumanEval, MBPP, or SWE-bench remains unverified — you're trusting Poolside's internal evals. The model likely struggles with architectural reasoning, test generation from specs, or explaining legacy code compared to Claude Sonnet 4.5 or o1. The XS size means less world knowledge for framework-specific edge cases or newer library APIs. If your task needs deep reasoning over code semantics rather than pattern completion, you'll hit the ceiling quickly.

Specifications

Provider
poolside
Category
llm
Context length
262,144 tokens
Max output
32,768 tokens
Modalities
text
License
proprietary
Released
2026-07-02

Pricing

Input
$0.06/Mtok
Output
$0.12/Mtok
Model ID
poolside/laguna-xs-2.1

Per-token prices show what the model costs upstream. On Switchy your team draws from one shared org credit pool - one plan, one balance for everyone.

Team cost calculator

Estimated monthly spend
$1.37
17.6M tokens / month
5 seats · 80 msgs/day

Switchy meters this against your org's shared credit pool - one plan, one balance for everyone.

Providers

ProviderContextInputOutputP50 latencyThroughput30d uptime
poolside262k$0.06/Mtok$0.12/Mtok

Performance

Performance snapshots are collected daily. Check back after the next ingestion run.

Benchmarks

Public benchmark scores are not available yet for this model. Check back after the next ingestion run.

Works well with

Top MCPs

Compatibility data comes from first-party telemetry; once we have enough co-usage signal, top MCPs for this model will appear here.

How Switchy teams use it

Not enough Spaces have used this model yet to share anonymised team stats. We wait for at least 50 distinct Spaces per week before publishing any aggregate.

Starter prompts

Autocomplete Function Body

Complete the following function implementation. The function signature and docstring are provided. Return only the function body:

```python
def calculate_percentile(data: list[float], percentile: int) -> float:
    """Calculate the nth percentile of a dataset."""
```
Open in a Space →

Generate Inline Docstring

Write a concise docstring for this function. Include parameter descriptions and return value:

```typescript
function debounce(func: Function, delay: number) {
  let timeoutId: NodeJS.Timeout;
  return function(...args: any[]) {
    clearTimeout(timeoutId);
    timeoutId = setTimeout(() => func.apply(this, args), delay);
  };
}
```
Open in a Space →

Explain Code Snippet

Explain what this code does in 2-3 sentences. Focus on the algorithm and edge cases:

```python
def longest_common_subsequence(s1: str, s2: str) -> int:
    m, n = len(s1), len(s2)
    dp = [[0] * (n + 1) for _ in range(m + 1)]
    for i in range(1, m + 1):
        for j in range(1, n + 1):
            if s1[i-1] == s2[j-1]:
                dp[i][j] = dp[i-1][j-1] + 1
            else:
                dp[i][j] = max(dp[i-1][j], dp[i][j-1])
    return dp[m][n]
```
Open in a Space →

Refactor for Readability

Refactor this function for readability. Improve variable names, add type hints, and simplify logic where possible:

```javascript
function p(a, b) {
  let r = [];
  for (let i = 0; i < a.length; i++) {
    if (a[i] > b) r.push(a[i]);
  }
  return r;
}
```
Open in a Space →

Find Potential Bugs

Review this code for potential bugs. List any issues with null handling, edge cases, or logic errors:

```python
def get_user_by_id(user_id: int, users: list[dict]) -> dict:
    for user in users:
        if user['id'] == user_id:
            return user
    return None
```
Open in a Space →
Data last verified 2 hours ago.Sources aggregated hourly to weekly. See docs/architecture/model-directory.md.