Laguna S 2.1
Unavailable in this round
Language modelActive
Laguna S 2.1
Poolside
- Released
- July 13, 2026
- Data date
- September 8, 2026
- Price
- Input · Output / 1M tokens
- Input
- Text
- Output
- Text
Laguna S 2.1 combines 118 billion total parameters with 8 billion active per token and a 1M-token context. Poolside achieves this with 12 global and 36 sliding-window attention layers, so not every step needs full global attention. It is the larger family variant of Laguna XS 2.1.
The model can think between tool calls, with that reasoning trace enabled or disabled per request through enable_thinking. OpenMDW-1.1 permits commercial use and modifications to the weights. Poolside says the BF16 checkpoint needs roughly 236 GB for its weights. Poolside also publishes FP8, NVFP4, INT4, and GGUF variants.
Specifications and access
| Specification | Value and source |
|---|---|
| Weights | Open weightsSource |
| Access | API, LocalSource |
| Parameters | 118BSource |
| Active parameters | 8BSource |
| Context window | 1,048,576 tokensSource |
| License | OpenMDW-1.1Source |
| Parameters | 118 BSource |
| Active parameters | 8 BSource |
| Architecture | sparse MoE, 48 layers (12 global and 36 sliding-window attention), 256 routed experts top-10 plus one shared expert, GQASource |
| Input | TextSource |
| Output | TextSource |
| Reasoning mode | interleaved thinking with preserved reasoning; per-request enable_thinking controlSource |
| Tool calling | YesSource |
Measurements without matching peer values
These measurements have no matching peer values under the same test conditions. Their original values and sources remain available here.
Terminal-Bench 2.1 (Show measurement, test conditions, and source)
- Source value
- 70.2
- Score
- 70.2
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026; exact sampling details in card.
SWE-bench Multilingual (Show measurement, test conditions, and source)
- Source value
- 78.5
- Score
- 78.5
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026.
SWE-Bench Pro (Public Dataset) (Show measurement, test conditions, and source)
- Source value
- 59.4
- Score
- 59.4
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026.
DeepSWE (Show measurement, test conditions, and source)
- Source value
- 40.4
- Score
- 40.4
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026.
SWE Atlas (Codebase QnA) (Show measurement, test conditions, and source)
- Source value
- 46.2
- Score
- 46.2
- Metric
- percent
- Unit
- %
- Category
- coding
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026.
Toolathlon Verified (Show measurement, test conditions, and source)
- Source value
- 49.7
- Score
- 49.7
- Metric
- percent
- Unit
- %
- Category
- agentic
- Direction
- Higher is better
- Source type
- vendor-reported
- Evaluator
- Poolside Laguna S 2.1 model card
- Status
- active
- Retrieved at
- 2026-09-08
- Methodology
- Vendor-reported result. Harbor Framework with Poolside agent harness; table dated July 21, 2026.
Our tests
SVG short test
One SVG prompt and fixed output settings. Captures are shown only with 0 reported reasoning tokens or documented lack of reasoning controls. Identical test conditions do not make the models' internal compute identical.
Earlier tests
Pelican on a bicycle
A community classic for free-form SVG drawing.
Laguna S 2.1
Reasoning (requested): high
Current attempt
- Run date
- Oct 1, 2026
- API provider
- openrouter
- Requested API model
- poolside/laguna-s-2.1
- API model reported in response
- poolside/laguna-s-2.1
- Reasoning setting
- high
- Pinned endpoint
- Poolside | poolside/laguna-s-2.1-20260720
- Temperature
- Provider default
- Token limit including reasoning
- 8,192 tokens
- Test protocol
- community-visual-tests-v1
- Technical checks
- Passed, not a quality rating
- Response time
- 16.2 s
Visually reviewed. This individual test does not establish overall model quality.
Prompt and test conditions
Generate an SVG of a pelican riding a bicycle
Token limit including reasoning: 8,192 tokens
Task origin (Simon Willison)Popularity
44.7 / 100
Sources and data date
| Type | Evidence and data date |
|---|---|
| Price | Base API prices without caching or batch discounts. Higher context tiers and other rates are listed under Costs. |
| Input | Documented modalities. Retrieved September 8, 2026. |
| Output | Documented modalities. Retrieved September 8, 2026. |
| Research status | Research date September 8, 2026. 2 source URLs checked. This documents the inspected sources, not an exhaustive inventory of every publication. |
| Not found in the inspected sources | knowledge-cutoff Original wording The exact model card does not publish a training-data cutoff. |
| Unresolved | AA benchmark rows Original wording No exact Artificial Analysis model page located in the checked public sitemap. |
| Additional source | OpenRouter: Popularity index (derived), CC BY 4.0 |
| Additional source | Vercel AI Gateway: Popularity index (derived), CC BY 4.0 |
| Additional source | Creative Commons Attribution 4.0 |