Laguna M.1
Laguna M.1A 225B total parameter Mixture-of-Experts model with 23B activated parameters per token designed for agentic coding and long-horizon work.
Model details
View repositoryExample usage
Laguna M.1 is a 225B total parameter Mixture-of-Experts model with 23B activated parameters per token designed for agentic coding and long-horizon work. Laguna M.1 uses global attention across all layers with 64 Q-heads, 8 KV-heads and softplus attention output gating. Apache 2.0 licensed.
This model was pre-trained on 30 trillion tokens and delivers the strongest performance in Poolside’s agent harness, pool, after undergoing agent RL. Poolside’s RL stack is a custom-built system loosely coupling the major components of inference and rollout generation, orchestration of code execution sandboxes, trajectory scoring, buffering and filtering, and distributed training.
Learn more here: https://poolside.ai/blog/introducing-laguna-xs2-m1.
Technical report: https://poolside.ai/assets/laguna/laguna-m1-xs2-technical-report.pdf
1# You can use this model with any of the OpenAI clients in any language!
2# Simply set the API Key to get started
3
4import os
5from openai import OpenAI
6
7model_url = "" # Copy in from API pane in Baseten model dashboard
8
9client = OpenAI(
10 api_key=os.environ['BASETEN_API_KEY'],
11 base_url=model_url,
12)
13
14response = client.chat.completions.create(
15 model="poolside/laguna-m.1",
16 messages=[
17 {
18 "role": "user",
19 "content": "What is the time complexity of quicksort?"
20 }
21 ],
22 extra_body={
23 "chat_template_kwargs": {"enable_thinking": False}
24 },
25 stream=False,
26)
27
28print(response.model_dump_json(indent=2))1{
2 "id": "chatcmpl-def456",
3 "object": "chat.completion",
4 "model": "poolside/laguna-m.1",
5 "choices": [
6 {
7 "index": 0,
8 "message": {
9 "role": "assistant",
10 "reasoning_content": null,
11 "content": "Quicksort has the following time complexities:\n\n- **Best case**: O(n log n) — pivot consistently splits the array into equal halves.\n- **Average case**: O(n log n) — expected with random pivot selection.\n- **Worst case**: O(n²) — occurs when the pivot is always the smallest or largest element (e.g. already-sorted input with naive pivot choice).\n\nSpace complexity is O(log n) on average for the call stack.",
12 "tool_calls": null
13 },
14 "finish_reason": "stop"
15 }
16 ],
17 "usage": {
18 "prompt_tokens": 15,
19 "completion_tokens": 95,
20 "total_tokens": 110
21 }
22}
