Runs on ≤32 GB · Hugging Face ↗
Not independently scored by Artificial Analysis
Poolside's first open-weight coder — 33B total / ~3B active MoE, Apache-2.0, built for agentic long-horizon coding. Reportedly 68.2% SWE-bench Verified and beating dense models 10x its size; the local-runnable member of the Laguna family (S 2.1 / M.1 are data-centre). Custom `LagunaForCausalLM` arch.
BigCodeBench-Hard pass@1 34% (41/121), via rented RTX PRO 6000 Blackwell 96 GB (vLLM, BF16) — comparable full-148, conc64. The brutal counterpart to HumanEval — where HumanEval saturates near the top, BCB-Hard spreads the field, so this is the number that actually separates coding ability. The local-runnable Laguna — lands with the strong open coders on hard problems; agentic by design, so a single-shot benchmark is a floor for its long-horizon SWE behaviour.
Real GGUF file sizes = weight VRAM. Add the KV cache for your context (≈5 GB at 32K, fp16).
| Quantization | Size |
|---|---|
| Q3_K_M | 15.6 GB |
| IQ4_XS | 18.2 GB |
| Q4_0 | 19.2 GB |
| Q4_K_S | 19.8 GB |
| Q4_K_M | 20.5 GB |
| Q4 | 20.7 GB |
| Q4_1 | 21.2 GB |
| Q5_K_M | 24.0 GB |
| Q2_K | 24.5 GB |
| Q8_0 | 35.6 GB |
| Q6_K | 58.2 GB |
| BF16 | 66.9 GB |
~20 GB at Q4 — a 24 GB+ card or the 32 GB tier. Also on OpenRouter (laguna-xs-2.1:free) and opencode Zen's free set, so you can try the family for $0 before committing local VRAM.