Laguna XS 2.1

# Fix list: Laguna XS 2.1

From Anchor Terminal's listing at https://www.anchorterminal.com/tools/laguna-xs-2-1, the October 2026 research run, assessed 21 October 2026. Grade C, 61 out of 100.

This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public.

For a coding agent working on Laguna XS 2.1: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published.

## 1. Risk, 55 out of 100, up to 13.8 more on the total

Why it scored 55: File safety 15 of 25, safetensors and GGUF only, but the repository ships model code behind auto_map and the card's SGLang and TensorRT-LLM commands use trust_remote_code. Provenance 15, the verified poolside organisation that Poolside's blog links, with a checksum on every file. Safety evaluation 0, none published. Agent guidance 5, general cautions only. Licence risk 15, OpenMDW-1.1 has no use policy, threshold or revocation clause. Security programme 5, a signed security.txt and a security contact on the card, with no public advisories found.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-risk):

What could go wrong when you run it, and what the publisher has done about it. A higher score means less risk.

- 0 to 25, file safety. 25 for weights in safetensors or GGUF only, 15 when pickle files (.bin, .pt, .pth) sit beside them, 0 for pickle only. 10 off when the documented way to load it or to build its prompts runs code from the repository (`trust_remote_code`, or the publisher's scripts in place of a template).
- 0 to 15, provenance. 10 when the files come from the publisher's own organisation, the one its site or docs link to, and 5 for a checksum on every file.
- 0 to 20, safety evaluation. 20 when results are published and someone other than the publisher ran them or can rerun them, 10 for the publisher's own results, 0 for none.
- 0 to 15, guidance for agent use. 15 for documented prompt-injection guidance or an instruction hierarchy the model was trained to follow, 10 for a statement of what its trained refusals cover and the safeguards to add around it, 5 for general cautions, 0 for none.
- 0 to 15, licence risk. 15 to start, 5 off for each of these, a use policy or field-of-use limits beyond the law, terms the licensor can change or revoke, a revenue or user threshold, and a licence that changed without a dated note or whose link is broken. Not below 0.
- 0 to 10, a security programme. 5 for a way to report a flaw (security.txt, a disclosure policy or a security contact), 5 for advisories handled in public.

When our reference-machine runs start, measured robustness (prompt injection with AgentDojo, over-refusal with XSTest) joins Risk.

## 2. Documentation and upkeep, 46 out of 100, up to 8.3 more on the total

Why it scored 46: Model card 11 of 20, intended use, architecture, parameters and context are there, limitations, languages and a cut-off aren't. Training data 10, the linked report describes sources and filtering and Poolside's S 2.1 post gives the shared November 2025 cut-off. Last change 15, 19 August 2026. Responsiveness 0, Poolside answered none of the 6 discussions. Version policy 5, XS.2 is kept and release notes are dated, but GGUF files were replaced in place without a card note. File health 5, generation config and template disagree on the thinking default and two file reports wait without a reply.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-upkeep):

- 0 to 20, the model card. Intended use, limitations, architecture, parameters, context length, languages and knowledge cut-off. 20 for all seven, 3 off for each that's missing.
- 0 to 20, training data. 20 for the data published or listed with sources and sizes, 15 for a data card that names the sources or the mix, the cut-off and the filtering, 10 for a general description with a cut-off, 5 without one, 0 for none.
- 0 to 20, the last change to the files, the template or the reference code. 20 within 30 days, 15 within 90, 10 within 180, 5 within a year, 0 older.
- 0 to 20, responsiveness. The publisher's replies to the 40 most recent questions and reports in the repository's discussions or issues. 20 for most, 10 for some, 5 for a few, 0 for none.
- 0 to 10, version policy. 10 for earlier versions kept in their own repositories with a dated note when one is replaced, 5 for earlier versions kept without notes or a first release, 0 when files were replaced in place without a note.
- 0 to 10, file health. 10 when config, tokenizer, template and card agree and no report of a broken file waits without a reply, 5 when one does, 0 when the files contradict the card.

## 3. Reach, 85 out of 100, up to 5.8 more on the total

Why it scored 85: Memory 20 of 30, the smallest 4-bit build with an exact size is Poolside's Q4_K_M GGUF at 20.27 GB, the same file as Ollama's default, which needs 32 GB with a fifth added. Runtimes 25, released llama.cpp, Ollama, vLLM, SGLang and Transformers all run the laguna architecture. Builds 15, Poolside publishes FP8, INT4, NVFP4, MLX and GGUF builds. Terms 10, OpenMDW-1.1 is permissive but not OSI-approved. Download 10, nothing is gated. Hosted 5, Poolside serves it on OpenRouter, though both endpoints are set to expire on 31 October 2026.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-reach):

Whether an agent can run it, and where.

- 0 to 30, memory. The smallest build from the publisher or a runtime's own library that keeps 4-bit precision or better, at its file size plus a fifth for the runtime and a working context. 30 for 16 GB or less, 25 for 24 GB, 20 for 32 GB, 15 for 64 GB, 10 for what fits our reference machine (the GPU share of 128 GB of unified memory), 5 for 256 GB, 0 above.
- 0 to 25, runtimes. 5 for each of llama.cpp, Ollama, vLLM, SGLang, LM Studio and Hugging Face Transformers that runs the architecture in a released version, from the runtime's own code, library or documentation. A publisher's fork or an unreleased branch doesn't count. 25 at most.
- 0 to 15, builds from the publisher. 15 for quantised builds at two precisions or more (FP8, 4-bit, GGUF), 10 for one, 0 when only other people's quantisations exist.
- 0 to 15, terms of use. 15 for an OSI-approved licence with nothing added, 10 for a custom licence that lets anyone use it commercially for free below a threshold or under a use policy, 5 for research or non-commercial use only, 0 for no licence or unclear terms.
- 0 to 10, download. 10 when the files download with no account and no terms to accept, 5 for a click-through that needs an account, 0 when a person approves each request.
- 5, hosted too. The publisher's own API serves it, or at least two inference providers do, so an agent without the hardware can still call it.

When our reference-machine runs start, speed on that machine (tokens per second and time to first token at the build that fits) joins Reach.

## 4. Agent interface, 85 out of 100, up to 2.3 more on the total

Why it scored 85: Chat template 20, a Jinja file in the repository. Tool-calling format 25, the card shows how tools go in and calls come back and names the poolside_v1 parser for three runtimes. Trained tool calling 20, the template renders several calls a turn. Structured output 0, not documented. Reasoning control 10, enable_thinking per request with the reasoning in reasoning_content. Context 10, 128K trained per Poolside's report, with 262,144 reached by YaRN.

The checklist (https://www.anchorterminal.com/benchmark/#checklist-interface):

What a model and a developer need to talk to it. Speech, image and other models with no chat interface read the tool-calling and structured-output lines for what the model takes and returns (voices, timestamps, streaming, typed outputs), and the note says how.

- 0 to 20, the chat template. 20 for a template in the repository (Jinja, in the tokenizer config or its own file), 10 when the prompt format is only given as code you run.
- 0 to 25, the tool-calling format. 25 when the publisher documents how tools are declared and how calls come back, with the parser to use in a named runtime, 10 when only the template or a runtime's own recipe says it.
- 0 to 20, tool calling the model was trained for. 20 with several calls in one turn, 10 for one call at a time.
- 0 to 15, structured output. 15 when JSON or JSON Schema output is documented for the model with how to ask for it, 5 when the card claims it without saying how.
- 10, reasoning control. Thinking that can be turned off or budgeted per request, with where the reasoning text comes back.
- 0 to 10, context. 10 for 128K tokens or more trained, 5 for 32K to 127K. Context reached only by scaling the position encoding counts one band down.

## What we couldn't check

What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it.

- Ollama's MLX nvfp4 tag is shown as 20 GB with no exact size. If it's under 20.0 GB it would need 24 GB and memory would rise to 25. It only runs on Apple Silicon
- unchecked, whether Poolside publishes security advisories on GitHub, since GitHub's web pages answered 403 and the API isn't enabled here
- Poolside's OpenRouter endpoints for XS 2.1 expire on 31 October 2026. After that, hosted rests on Poolside's own API, which needs a key to list models
- The card calls XS 2.1 an upgraded version of Laguna XS.2 without saying it was trained from it, so no builtOn is set

## Weaknesses

- No reply from Poolside to any discussion, including reports about the tokenizer and remote code
- No safety results and only general cautions for agent use
- The SGLang and TensorRT-LLM commands run the repository's Python with trust_remote_code
- The template defaults thinking off while generation_config.json turns it on
- No structured output documentation, and no limitations, languages or cut-off on the card

## What costs an agent a turn today

The notes we give agents before they call it. Each one is a workaround an agent shouldn't need.

- Quickest start is `ollama run laguna-xs-2.1`, 20 GB of files, the same Q4_K_M as Poolside's GGUF
- On our reference machine use `llama-server -hf poolside/Laguna-XS-2.1-GGUF:Q4_K_M --jinja -c 131072`, from a llama.cpp release with the laguna architecture (b11772 has it), or BF16 with vLLM (66.9 GB)
- With vLLM use `--tool-call-parser poolside_v1 --reasoning-parser poolside_v1 --enable-auto-tool-choice`, on a build with vllm-project/vllm#47311, or `--tool-call-parser glm47` on older builds
- Thinking is off in the template unless you pass `enable_thinking` true or start the server with `--default-chat-template-kwargs '{"enable_thinking": true}'`, which Poolside recommends for agentic coding
- Pass earlier reasoning back as `reasoning_content` in the history, since the card says reasoning may stop in later steps if it's dropped
- Sampling per the card is temperature 1.0 and top_k 20
- The card warns that Ollama on Apple Silicon may return empty output until llama.cpp#25389 ships
- Validate JSON output yourself or use the runtime's guided decoding, since structured output isn't documented

## What the review panel asked for

- note rebuilds on the card
- a hosting plan after 31 October
- a JSON Schema output guide
- publish safety results

## When it's done

Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.