Tiiny Pocket verification
Imported from this repo onto a Tiiny Pocket (TiinyOS 0.1.39) and tested on 2026-11-17:
- Result: pass, 5/5 smoke tests passed
- NPU share 47%, listed size 72.1 GB (the device lists the source size)
- At 3719 prompt tokens: TTFT 4.83 s, prefill 770.1 tok/s, decode 26.8 tok/s
Full report: tiiny-verification.md (raw data in tiiny-verification.json).
Repackaged for Tiiny Pocket import. A copy of ValiantLabs/Qwen3.6-35B-A3B-Esper3.1 at d53eef4563a2, made by tiiny-forge so the Tiiny's Hugging Face import accepts it as a fine-tune of the store model Qwen/Qwen3.6-35B-A3B. The weights are unchanged. Changes:
- config.json: the base's (Qwen/Qwen3.6-35B-A3B), same tensors, its transformers schema
Support our open-source dataset and model releases!

Esper 3.1: Ministral-3-3B-Reasoning-2512, Qwen3-4B-Thinking-2507, Ministral-3-8B-Reasoning-2512, Ministral-3-14B-Reasoning-2512, gpt-oss-20b, Qwen3.5-27B, Qwen3.6-27B, Qwen3.6-35B-A3B
Esper 3.1 is a coding, architecture, and DevOps reasoning specialist built on Qwen 3.6 35B A3B!
- Your dedicated DevOps expert: Esper 3.1 maximizes DevOps and architecture helpfulness, powered by high-difficulty DevOps and architecture data generated with DeepSeek-V3.1-Terminus!
- Improved coding performance: challenging code-reasoning datasets stretch DeepSeek-V3.1-Terminus and DeepSeek-V3.2 to the limits, allowing Esper 3.1 to tackle harder coding tasks!
- AI to build AI: our high-difficulty AI expertise data boosts Esper 3.1's MLOps, AI architecture, AI research, and general reasoning skills.
- Small model sizes allow running on local desktop and mobile, plus super-fast server inference!
Prompting Guide
Esper 3.1 uses the Qwen3.6-35B-A3B prompt format.
Example inference script to get started:
from transformers import AutoModelForCausalLM, AutoTokenizer model_name = "ValiantLabs/Qwen3.6-35B-A3B-Esper3.1" # load the tokenizer and the modeltokenizer = AutoTokenizer.from_pretrained(model_name)model = AutoModelForCausalLM.from_pretrained( model_name, torch_dtype="auto", device_map="auto") # prepare the model inputprompt = "Write a Terraform configuration that uses the `aws_ami` data source to find the latest Amazon Linux 2 AMI. Then, provision an EC2 instance using this dynamically determined AMI ID."messages = [ {"role": "user", "content": prompt}]text = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True, enable_thinking=True # Switches between thinking and non-thinking modes. Default is True.)model_inputs = tokenizer([text], return_tensors="pt").to(model.device) # conduct text completiongenerated_ids = model.generate( **model_inputs, max_new_tokens=32768)output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() # parsing thinking contenttry: # rindex finding 248069 (</think>) index = len(output_ids) - output_ids[::-1].index(248069)except ValueError: index = 0 thinking_content = tokenizer.decode(output_ids[:index], skip_special_tokens=True).strip("\n")content = tokenizer.decode(output_ids[index:], skip_special_tokens=True).strip("\n") print("thinking content:", thinking_content)print("content:", content)

Esper 3.1 is created by Valiant Labs.
Check out our HuggingFace page to see all of our models!
We care about open source. For everyone to use.