LocalFit
Submit a benchmark

Laguna-XS-2.1 Coding

poolside/Laguna-XS-2.1

ReleasedRelease date unknownSize33B MoE · 3B activeRoleCodingLicenseopenmdw-1.1Context256K config max

Where it runs 8K context · ≥ 5 t/s floor

SetupBest planMemoryGenerate
8 GB GPU + 32 GB RAMQ4_K_Mllama.cpp CUDAGPU + CPU experts · 28 expert layers on CPUFits~13–22 t/s Estimated extrapolated
12 GB GPU + 32 GB RAMQ4_K_Mllama.cpp CUDAGPU + CPU experts · 19 expert layers on CPUFits~12–19 t/s Estimated extrapolated
16 GB GPU + 64 GB RAMQ4_K_Mllama.cpp CUDAGPU + CPU experts · 11 expert layers on CPUFits~33–55 t/s Estimated extrapolated
24 GB GPU + 64 GB RAMQ4_K_Mllama.cpp CUDAFull GPUFits~92–265 t/s Estimated extrapolated
32 GB GPU + 128 GB RAMQ4_K_Mllama.cpp CUDAFull GPUFits~163–472 t/s Estimated extrapolated
2× 24 GB GPUs + 128 GB RAMQ4_K_Mllama.cpp CUDAFull GPUFits~85–246 t/s Estimated extrapolated
96 GB workstation GPU + 256 GB RAMQ4_K_Mllama.cpp CUDAFull GPUFits~163–472 t/s Estimated extrapolated
Mac, 24 GBDoes not fit
Mac, 48 GBQ4_K_Mllama.cpp MetalUnified memoryFits~25–72 t/s Estimated extrapolated
Mac, 128 GBQ4_K_Mllama.cpp MetalUnified memoryFits~50–144 t/s Estimated extrapolated
Mac, 512 GBQ4_K_Mllama.cpp MetalUnified memoryFits~75–216 t/s Estimated extrapolated
DGX Spark 128 GBQ4_K_Mllama.cpp CUDAUnified memoryFits~40–107 t/s Estimated
Ryzen AI Max+ 395, 128 GBQ4_K_Mllama.cpp VulkanUnified memoryFits~23–67 t/s Estimated extrapolated
CPU only, 64 GB DDR5-5600 dual channelQ4_K_Mllama.cpp CPUCPU onlyFits~9.8–12 t/s Estimated extrapolated

Same plan engine as the planner, with its default reserves and RAM. How plans are chosen.

Files exact Hugging Face bytes

ArtifactSizeFilesRevisionSource
Q4_K_M
poolside/Laguna-XS-2.1-GGUF
18.88 GiB11a37c0a5fbHF files
Native safetensors BF16
poolside/Laguna-XS-2.1
62.30 GiB14c5f36269bbHF files

Not found in the representative quant search (not proven unavailable): GGUF FP16, GGUF Q8_0, GGUF Q5_K_M, MLX 8-bit, MLX 4-bit, EXL3, AWQ, GPTQ. Complete selected file sets; projectors, MTP/draft and lookup files are listed separately and budgeted only by plans that use them. File bytes are not allocated memory.

Quality

No Artificial Analysis evaluation mapped to this checkpoint.

Arena text rating

Unknown

Publisher-reported benchmarks

None extracted from the model card.

AA and Arena describe the hosted model in the stated mode, not each quantized file; scores are never inherited by base, pruned, abliterated or quantized variants and never averaged together.

How to run it

vLLM

source-documented, not executed source · vllm>=0.21.0

pip install 'vllm>=0.21.0'

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

llama.cpp fork/PR

source-documented, not executed source · PR25165 branch perpublisher

# Build llama.cpp from the PR branch
git clone https://github.com/ggml-org/llama.cpp && cd llama.cpp
git fetch origin pull/25165/head:laguna && git checkout laguna
cmake -B build && cmake --build build -j

# Download a GGUF and serve an OpenAI-compatible endpoint
huggingface-cli download poolside/Laguna-XS-2.1-GGUF Laguna-XS-2.1-Q4_K_M.gguf --local-dir ~/models/Laguna-XS-2.1-GGUF
./build/bin/llama-server -m ~/models/Laguna-XS-2.1-GGUF/Laguna-XS-2.1-Q4_K_M.gguf --jinja --port 8000

Recipes are source examples (settings are the publisher’s or author’s, not measured optima) and were not executed by LocalFit. A downloadable file does not prove your backend build supports the architecture.

Reported speeds 0

No reported measurement yet. Submit one.