Modelpublic

Accio-Lab/occamy-1.0

sourceHugging Faceapache-2.0updated 6d agoView on Hugging Face

183likes2kdownloads

Model Card

<div align="center"> <picture> <img src="assets/accio.svg" width="34%" alt="Accio"> </picture>      <picture> <img src="assets/occamy.png" width="13%" alt="Occamy logo"> </picture> <h1>Occamy-1.0</h1> Open Pareto-frontier 35B Intelligence for Co-work </div>

<hr>

<a href="https://accio-lab.github.io/occamy/">Project Website</a>  |  <a href="https://huggingface.co/Accio-Lab/Occamy-1.0">Model Weights</a>  |  <a href="https://github.com/Accio-Lab/Dressage">Training Framework</a>

Available Checkpoints

Format	Download
GGUF (Q4KM / Q8_0)	occamy-1.0-GGUF
FP8	occamy-1.0-FP8
NVFP4	occamy-1.0-NVFP4
Experimental MTP head	occamy-1.0-MTP

See each model card for usage and validation. Community releases also offer smaller GGUF quantizations and MLX builds for Apple Silicon.

BF16, FP8 and NVFP4 each passed a nine-case vLLM compatibility check on a single H200, covering text, code, JSON, tool calls and follow-up, and images. NVFP4 used Marlin W4A16. These checks used ordinary decoding; the separate MTP head remains experimental.

1. Model Introduction

Occamy-1.0 is a compact agentic model purpose-built for real-world co-work: long-horizon, stateful tasks that require coordinated use of search, code, tools, files, structured APIs, and productivity software. Starting from the post-trained Qwen3.6-35B-A3B checkpoint, Occamy concentrates further training on reliable execution, persistent state tracking, recovery, and follow-through rather than relearning general capabilities from scratch.

Key Features

—Co-work specialization: Designed for sustained execution across multi-step professional workflows, not isolated question answering.
—Compact inference footprint: A 35B-total, 3B-active Mixture-of-Experts model that keeps long-running agent workloads practical.
—Long-horizon continuity: Designed to keep work coherent across tool calls, delegated runs, and history rewrites such as context compaction.
—Broad agentic capability: Co-work gains are accompanied by strong tool calling, terminal coding, and instruction following.
—Execution-grounded training: Supervised fine-tuning spans general agentic work, long-horizon interaction, software engineering, and tool-call grounding.
—Open training stack: The multi-harness reinforcement-learning infrastructure used to train Occamy is released as Dressage.

[!NOTE] Occamy is optimized for common co-work workloads, not as a replacement for frontier models on every task. Retrieval-heavy and simulated-user tasks still have headroom, and native browser or desktop visual interaction is not part of the current co-work training interface.

2. Model Summary

<div align="center"> <table> <tbody> <tr><td align="center">Architecture</td><td align="center">Mixture-of-Experts causal model with vision encoder</td></tr> <tr><td align="center">Total Parameters</td><td align="center">35B</td></tr> <tr><td align="center">Activated Parameters</td><td align="center">3B</td></tr> <tr><td align="center">Number of Layers</td><td align="center">40</td></tr> <tr><td align="center">Number of Experts</td><td align="center">256</td></tr> <tr><td align="center">Activated Experts</td><td align="center">8 routed + 1 shared</td></tr> <tr><td align="center">Base Architecture Context</td><td align="center">262,144 tokens</td></tr> <tr><td align="center">SFT Sequence Length</td><td align="center">131,072 tokens</td></tr> <tr><td align="center">Starting Checkpoint</td><td align="center"><a href="https://huggingface.co/Qwen/Qwen3.6-35B-A3B">Qwen3.6-35B-A3B</a></td></tr> <tr><td align="center">Post-training</td><td align="center">Full-parameter SFT, HDPO, model merging, and SAO</td></tr> </tbody> </table> </div>

Architecture fields follow the starting checkpoint's published model card. Occamy post-trains the language backbone without changing the architecture; the vision encoder and projector are frozen during SFT. The released checkpoint configuration remains the source of truth for serving limits.

3. Evaluation Results

Full Evaluation

<table> <thead> <tr> <th rowspan="2" align="left" bgcolor="#EAF6F3">Benchmark</th> <th colspan="6" align="center" bgcolor="#EAF6F3">35B-A3B Models</th> <th colspan="4" align="center" bgcolor="#EAF6F3">Large-scale Models</th> </tr> <tr> <th align="center" bgcolor="#087F6A">Occamy-1.0</th> <th align="center">Qwen3.6 35B-A3B</th> <th align="center">Agents-A1</th> <th align="center">Nex-N2-mini</th> <th align="center">BigBang-1.0</th> <th align="center">Ornith-1.5</th> <th align="center">GPT-5.6 Sol</th> <th align="center">Qwen3.8-Max</th> <th align="center">DeepSeek V4 Pro (0813)</th> <th align="center">GLM-5.2</th> </tr> </thead> <tbody> <tr><td colspan="11" bgcolor="#DDF1EC">Co-work</td></tr> <tr><td>Claw-Eval (average)</td><td align="right">82.20</td><td align="right">69.50</td><td align="right">69.90</td><td align="right">66.60</td><td align="right">63.50</td><td align="right">64.40</td><td align="right">81.80</td><td align="right">83.92</td><td align="right">81.70</td><td align="right">81.60</td></tr> <tr><td>Claw-Eval (Pass³)</td><td align="right">71.40</td><td align="right">54.80</td><td align="right">41.70</td><td align="right">37.00</td><td align="right">40.20</td><td align="right">48.70</td><td align="right">68.90</td><td align="right">73.68</td><td align="right">74.50</td><td align="right">68.30</td></tr> <tr><td>WildClawBench</td><td align="right">49.16</td><td align="right">40.40</td><td align="right">30.73</td><td align="right">30.31</td><td align="right">32.87</td><td align="right">45.91</td><td align="right">67.20</td><td align="right">54.42</td><td align="right">37.30</td><td align="right">52.14</td></tr> <tr><td>CommerceAgentBench</td><td align="right">37.38</td><td align="right">19.60</td><td align="right">9.30</td><td align="right">16.80</td><td align="right">30.80</td><td align="right">37.40</td><td align="right">49.50</td><td align="right">46.30</td><td align="right">43.30</td><td align="right">39.30</td></tr> <tr><td>Business Arena</td><td align="right">$79,868</td><td align="right">$44,751</td><td align="right">$33,626</td><td align="right">$13,325</td><td align="right">$56,477</td><td align="right">$66,292</td><td align="right">$168,867</td><td align="right">$89,423</td><td align="right">$40,804</td><td align="right">$55,742</td></tr> <tr><td>GDPval†</td><td align="right">1,128</td><td align="right">1,004</td><td align="right">869</td><td align="right">999</td><td align="right">951</td><td align="right">855</td><td align="right">1,741</td><td align="right">1,640</td><td align="right">1,500</td><td align="right">1,452</td></tr> <tr><td>OfficeQA Pro</td><td align="right">48.10</td><td align="right">39.10</td><td align="right">23.30</td><td align="right">46.60</td><td align="right">43.60</td><td align="right">59.40</td><td align="right">74.40</td><td align="right">69.20</td><td align="right">51.20</td><td align="right">66.20</td></tr> <tr><td>τ³-Bench (Banking)</td><td align="right">37.10</td><td align="right">11.90</td><td align="right">7.20</td><td align="right">25.80</td><td align="right">10.30</td><td align="right">21.70</td><td align="right">46.90</td><td align="right">54.60</td><td align="right">44.30</td><td align="right">37.10</td></tr> <tr><td colspan="11" bgcolor="#DDF1EC">Tool calling</td></tr> <tr><td>AutomationBench (Pass¹)</td><td align="right">27.60</td><td align="right">7.50</td><td align="right">2.20</td><td align="right">5.70</td><td align="right">14.80</td><td align="right">18.50</td><td align="right">45.50</td><td align="right">43.50</td><td align="right">32.00</td><td align="right">28.00</td></tr> <tr><td>AutomationBench (partial)</td><td align="right">69.10</td><td align="right">39.40</td><td align="right">14.70</td><td align="right">27.90</td><td align="right">47.40</td><td align="right">58.00</td><td align="right">81.20</td><td align="right">81.20</td><td align="right">59.70</td><td align="right">70.00</td></tr> <tr><td>BFCL v4</td><td align="right">65.40</td><td align="right">63.19</td><td align="right">57.23</td><td align="right">62.81</td><td align="right">57.86</td><td align="right">68.51</td><td align="right">64.33</td><td align="right">73.65</td><td align="right">67.10</td><td align="right">70.33</td></tr> <tr><td>VitaBench</td><td align="right">41.75</td><td align="right">34.25</td><td align="right">37.00</td><td align="right">26.25</td><td align="right">46.00</td><td align="right">40.25</td><td align="right">46.75</td><td align="right">52.25</td><td align="right">53.50</td><td align="right">43.75</td></tr> <tr><td colspan="11" bgcolor="#DDF1EC">Coding</td></tr> <tr><td>Terminal-Bench 2.1</td><td align="right">59.00</td><td align="right">49.50</td><td align="right">41.60</td><td align="right">60.70</td><td align="right">33.70</td><td align="right">67.80</td><td align="right">88.80</td><td align="right">81.30</td><td align="right">87.90</td><td align="right">82.70</td></tr> <tr><td colspan="11" bgcolor="#DDF1EC">Instruction following</td></tr> <tr><td>IFEval</td><td align="right">91.53</td><td align="right">86.90</td><td align="right">91.60</td><td align="right">91.60</td><td align="right">90.50</td><td align="right">81.80</td><td align="right">95.00</td><td align="right">95.02</td><td align="right">93.74</td><td align="right">93.89</td></tr> </tbody> </table>

Within each size group, bold denotes the best result and underlining denotes the second-best result. * Official model-card or Artificial Analysis result. † Reproduced on the public task release.

Cost-Performance

Across Claw-Eval, WildClawBench, AutomationBench, and GDPval, Occamy-1.0 lies near the low-cost knee of the empirical Pareto frontier. Relative to its Qwen3.6-35B-A3B starting checkpoint, it delivers a large aggregate capability gain with only a modest change in measured per-task inference cost. Benchmark scores are equally weighted after per-benchmark min-max normalization, and costs are macro-averaged per task under the frozen pricing protocol used in the report.

4. Training Recipe

Occamy uses staged specialization and consolidation:

text

Qwen3.6-35B-A3B
  ├─ Marathon Expert: SFT → HDPO ┐
  └─ Sprint Expert: SFT          ├─ Uniform merge → SAO → Occamy-1.0

The Marathon Expert learns sustained execution and accuracy-conditioned efficiency, while the Sprint Expert preserves broader agentic capability. A uniform parameter-space merge combines both experts into one checkpoint with no inference-time routing or ensembling, and a final Single-Rollout Asynchronous Optimization (SAO) stage refines the merged policy on a broad co-work mixture.

The deduplicated SFT union across both experts is:

Data source	Trajectories	Average length	Tokens
General agentic	5,418	37.7K	204.1M
Long-horizon interactive agents	923	95.8K	88.4M
Terminal and software engineering	1,228	35.1K	43.1M
Tool-call grounding	7,429	9.1K	67.7M
Overall	14,998	26.9K	403.3M

Training tasks are grounded in executable environments with observable state transitions and task-level grading. The open-source Dressage stack provides multi-harness execution, token-exact trajectory capture, sandbox integration, and multi-segment conversion for reinforcement learning.

5. Deployment

Occamy-1.0 keeps the Qwen3.6-35B-A3B architecture, so the upstream deployment recipe is the reference serving path. The examples below mirror that recipe with eight-way tensor parallelism and its full context length; adjust both to fit your hardware and confirm them against the released Occamy checkpoint configuration.

SGLang

The upstream model card recommends SGLang 0.5.10 or newer for the Qwen3.6 architecture.

bash

python -m sglang.launch_server \
  --model-path Accio-Lab/Occamy-1.0 \
  --port 8000 \
  --tp-size 8 \
  --mem-fraction-static 0.8 \
  --context-length 262144 \
  --reasoning-parser qwen3 \
  --tool-call-parser qwen3_coder

vLLM

The upstream model card recommends vLLM 0.19.0 or newer for the Qwen3.6 architecture.

bash

vllm serve Accio-Lab/Occamy-1.0 \
  --port 8000 \
  --tensor-parallel-size 8 \
  --max-model-len 262144 \
  --reasoning-parser qwen3 \
  --enable-auto-tool-choice \
  --tool-call-parser qwen3_coder

Both commands expose an OpenAI-compatible endpoint at http://localhost:8000/v1.

6. Model Usage

python

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")

response = client.chat.completions.create(
    model="Accio-Lab/Occamy-1.0",
    messages=[
        {
            "role": "user",
            "content": "Inspect this repository, fix the failing test, and explain the change.",
        }
    ],
    max_tokens=32768,
    temperature=1.0,
    top_p=0.95,
    presence_penalty=1.5,
    extra_body={
        "top_k": 20,
        "chat_template_kwargs": {
            "enable_thinking": True,
            "preserve_thinking": True,
        },
    },
)

print(response.choices[0].message.content)

For multi-turn agent runs, retain the complete assistant message returned by the server, including reasoning content and tool calls, then append tool results using the standard OpenAI chat-completions schema. This preserves the execution context that Occamy relies on across long workflows.

Agent Frameworks

Occamy was trained and evaluated across multiple harnesses, including OpenClaw, Hermes Agent, and Accio Work. It can be integrated with other tool-using agent frameworks through the same OpenAI-compatible API.

7. License

This repository is released under the Apache License 2.0. See the Hugging Face model card for the terms that apply to the model weights.

8. Contact Us

For questions or feedback, please open an issue.