The 33B model that just beat 137B models on coding benchmarks without changing hardware

Poolside released Laguna XS 2.1 today, an incremental update to its coding-focused Mixture-of-Experts model. The headline number is a 5.4-point jump on SWE-bench Multilingual, bringing the resolved task rate to 63.1%. That sounds like a minor improvement until you look at the company's benchmark table: XS 2.1 now beats models with 10 to 40 times more active parameters on several agentic coding tests. The pattern of efficient local models challenging datacenter-scale deployments has been building for two years, and Poolside is the latest to make it work, as the full benchmark comparison shows.

The model keeps the same architecture as XS.2, 33B total parameters, 3B activated per token, but the company says it trained longer or on better data, or both. The release post does not specify what changed in training, only that the result is "a notable improvement on SWE-bench Multilingual and stronger performance on terminal-style tasks." That vagueness is common in model release notes, but it matters here because the benchmark improvements come with no hardware cost increase. XS 2.1 costs the same on Poolside's API as XS.2: $0.10 per million input tokens, $0.20 per million output tokens, $0.05 per million cache-read tokens. For teams already running XS.2 locally on a single GPU, the upgrade is essentially free compute-wise.

What the benchmarks actually show

Poolside published scores across four agentic coding benchmarks: SWE-bench Verified, SWE-bench Multilingual, SWE-Bench Pro, and Terminal-Bench 2.0. The company compared XS 2.1 against Qwen3.6 (35B-A3B), North Mini Code from Cohere (30B-A3B), MAI-Code-1-Flash (137B-A5B), GPT-OSS-120B, and Claude Haiku 4.5. On SWE-bench Multilingual, XS 2.1 leads the comparison set at 63.1%. On SWE-bench Verified, it trails several larger models. On Terminal-Bench 2.0, it roughly ties Qwen3.6.

The takeaway is not that XS 2.1 dominates every benchmark, it doesn't. Across four tests, it leads one, roughly ties one, and trails on two. But it does all of this with only 3B active parameters. A model that fits on a single consumer GPU competes with offerings that need datacenter hardware, and that changes the economics of local AI coding for individual developers and small teams. The broader implications for agent pricing are explored in Nvidia's analysis on agent cost savings.

Speculative decoding and local inference

Poolside is also releasing DFlash speculator models for each XS 2.1 checkpoint. These are small draft models trained to predict the main model's next tokens cheaply, a technique called speculative decoding. In the company's internal tests, pairing XS 2.1 with its speculator doubles the achieved tokens per second. For anyone running the model locally, where latency is the bottleneck, that makes a meaningful difference in how responsive the coding agent feels. This technique is described in more detail by the mechanics of speculative decoding.

The model supports vLLM, SGLang, NVIDIA TensorRT-LLM, Hugging Face transformers, and Ollama. Quantized checkpoints are available in FP8, INT4, and NVFP4, letting users trade precision for VRAM. The company says quantized GGUF checkpoints and llama.cpp support are coming soon.

License change and ecosystem play

Poolside is licensing XS 2.1 under OpenMDW-1.1, dropping whatever license XS.2 used. OpenMDW is the open model distribution framework backed by NVIDIA and the Linux Foundation. The company calls it "fully permissive" and says the change "reduc[es] licensing friction for open model releases." For teams evaluating models, this removes one more headache, no need to run a license review for a model that might end up in a commercial product.

The timing aligns with a broader push from NVIDIA to standardize how open models are distributed. Poolside's story is not just about benchmark numbers; it is about making those numbers accessible on local hardware under a license that lawyers will sign off on.

What gets deprioritized

Poolside will sunset XS.2 on its own API after one week. Teams that built dedicated deployments around the older model can still access it through Baseten's Model Library, but the company is clearly pushing everyone toward 2.1. The one-week sunset window on the API is short, teams running evaluations or regression tests will need to move quickly.

The company also made a deliberate bet on agentic coding over general-purpose reasoning. XS 2.1 is designed for "long-horizon work on a local machine," meaning the kind of multi-step coding tasks that an agentic loop handles. It is not a generalist chatbot. Teams using it for anything other than code may find it narrower than the benchmark table suggests.

The bigger picture

Poolside's release fits a pattern that has been building for two years: small, efficient models that run on local hardware are eating into the territory that once belonged exclusively to massive datacenter deployments. XS 2.1 is not a breakthrough, it is an incremental improvement that makes a good argument slightly better. The 5.4-point gain on one benchmark is modest. What matters is that Poolside achieved it without increasing the model's footprint, the API pricing, or the hardware requirements. That is the kind of improvement that compounds over time, and it is how local AI coding becomes a default workflow rather than a curiosity. For those still skeptical of local model viability, the case for local AI remains compelling.