TobDeBer/Laguna-S-2.1-GGUF

Model details

This model is a hard requant of Laguna-S-2.1-UD-IQ3_XXS.gguf.
A Big Thank You goes out to the unsloth team for providing that!

To run it you need a llama.cpp build of 2026-07-23 or later!

... but why?!?!

This is a hard compression of a 118b model to below 20GB.
This means it destroys a lot of the capabilities of the original model!

So why the heck did I do this? As an experiment if such a huge model compressed hard is still better than a more right-sized model at the 24GB GPU tier.

Let's see how it turns out. Please add your results in the comments section.

Running with llama.cpp

The GGUFs in this repo are hard requants based on Unsloth Dynamic 2.0 quants (imatrix calibrated).

Download the UD-IQ3_XXS_Q1exps (~19GB):

huggingface-cli download TobDeBer/Laguna-S-2.1-GGUF \
  --include "Laguna-S-2.1-UD-IQ3_XXS_Q1exps.gguf" \
  --local-dir Laguna-S-2.1-GGUF

Serve it with llama-server:

./llama.cpp/build/bin/llama-server \
    --model Laguna-S-2.1-GGUF/Laguna-S-2.1-UD-IQ3_XXS_Q1exps.gguf \
    --jinja -fa on -ngl 99 --ctx-size 16384 --port 8000

Or run a one-off generation with llama-cli:

./llama.cpp/build/bin/llama-cli \
    --model Laguna-S-2.1-GGUF/Laguna-S-2.1-UD-IQ3_XXS_Q1exps.gguf \
    --jinja -ngl 99 -p "Write a Flappy Bird game in Python."

-ngl 99 offloads all layers to GPU; lower it (or drop it) if you run out of VRAM.

run with DFlash speculative decoding (download DFlash.ggug first):

  -md laguna-s-2.1-DFlash-BF16.gguf \
  --spec-type draft-dflash --spec-draft-n-max 15 -fa on --jinja --port 8000

Use on OpenRouter · Use on Vercel AI Gateway · Release blog post

Laguna S 2.1

Laguna S 2.1 is a 118B total parameter Mixture-of-Experts model with 8B activated parameters per token, designed for agentic coding and long-horizon work. It sits between Laguna XS 2.1 (33B-A3B) and Laguna M.1 (225B-A23B) in the Laguna series and shares the family recipe: a token-choice router with softplus gating over 256 routed experts plus one shared expert, grouped-query attention, and interleaved full/sliding-window attention.

Highlights

• Mixed SWA and global attention layout: 48 layers in a 1:3 global-to-SWA ratio (12 global attention layers, 36 sliding-window layers, window 512), with softplus attention gating and per-layer-type rotary scales
• 1M context: 1,048,576-token context window
• Native reasoning support: interleaved thinking between tool calls, with per-request control via enable_thinking
• Speculative decoding: a trained DFlash draft model is available for lower-latency serving
• Quantized variants: FP8, NVFP4, INT4 and GGUF
• OpenMDW-1.1 license: Use and modify the model and associated materials freely for commercial and non-commercial purposes (learn more about OpenMDW)

Model overview

• Number of parameters: 118B total, ~8B activated per token
• Layers: 48 (12 global attention, 36 sliding-window attention)
• Experts: 256 routed (top-10) plus 1 shared expert
• Attention: grouped-query, 8 KV heads, head dim 128; per-head softplus output gating
• Sliding window: 512 tokens
• Context window: 1,048,576 tokens
• Vocabulary: 100,352 tokens (Laguna family tokenizer)
• Modality: text-to-text
• Reasoning: interleaved thinking with preserved thinking

Benchmark results

Model	Size	Terminal-Bench 2.1	SWE-bench Multilingual	SWE-Bench Pro (Public Dataset)	DeepSWE	SWE Atlas (Codebase QnA)	Toolathlon Verified
Laguna S 2.1	118B-A8B	70.2%	78.5%	59.4%	40.4%	46.2%	49.7%
Tencent Hy3	295B-A21B	71.7%	75.8%	57.9%	-	-	-
Inkling	975B-A41B	63.8%	-	54.3%	-	-	45.5%*
Nemotron 3 Ultra	550B-A55B	56.4%	67.7%	-	-	-	34.3%*
DeepSeek-V4-Pro Max	1.6T-A49B	64.0%*	76.2%	55.4%	9.0%*	27.2%*	55.9%*
Kimi K3	2800B-A50B	88.3%	-	-	69%	-	-
Qwen 3.7 Max	-	74.5%*	78.3%	60.6%	-	-	-
Muse Spark 1.1	-	80%	-	61.5%	53.3%	42.2%*	75.6%
Claude Fable 5	-	88%	-	80.3%	70%	-	-

Benchmarks as of 21 July 2026. Laguna S 2.1 in bold; a dash (-) marks a benchmark a model was not evaluated on. Scores marked * are as reported by third parties: Terminal-Bench 2.1 and DeepSWE via Artificial Analysis, SWE Atlas via Scale AI's official leaderboard, and Toolathlon Verified via its official leaderboard. Full evaluation trajectories: trajectories.poolside.ai.

Usage

Laguna S 2.1 uses the same laguna architecture as Laguna XS 2.1, so the same engine integrations apply (vLLM, SGLang, Transformers, TRT-LLM, llama.cpp). At 118B parameters the BF16 checkpoint needs multiple GPUs (roughly 236GB of weights); quantized variants reduce this substantially.

vLLM

vllm serve \
    --model poolside/Laguna-S-2.1 \
    --tensor-parallel-size 4 \
    --tool-call-parser poolside_v1 \
    --reasoning-parser poolside_v1 \
    --enable-auto-tool-choice \
    --served-model-name laguna \
    --default-chat-template-kwargs '{"enable_thinking": true}'

Optional: speculative decoding with DFlash. Pair with the Laguna S 2.1 DFlash draft model by adding --speculative-config '{"model":"poolside/Laguna-S-2.1-DFlash","num_speculative_tokens":7,"method":"dflash"}'.

Controlling reasoning

Laguna S 2.1 has native reasoning support and works best with preserved thinking: keep reasoning_content from prior assistant messages in the message history. The model will generally reason before calling tools and between tool calls, and may stop reasoning in follow-up steps if prior thinking blocks are dropped.

Thinking is controlled per request via the chat template:

extra_body={"chat_template_kwargs": {"enable_thinking": False}}

or at the server level with --default-chat-template-kwargs '{"enable_thinking": true}'. For agentic coding use cases we recommend enabling thinking and preserving reasoning in the message history.

License

This model is licensed under the OpenMDW-1.1 License.

Intended and Responsible Use

Laguna S 2.1 is designed for software engineering and agentic coding use cases, and you are responsible for confirming that it is appropriate for your intended application. Laguna S 2.1 is subject to the OpenMDW-1.1 License, and should be used consistently with Poolside's Acceptable Use Policy. We advise against circumventing Laguna S 2.1 safety guardrails without implementing substantially equivalent mitigations appropriate for your use case.

Please report security vulnerabilities or safety concerns to security@poolside.ai.