Ornith 1.0 35B

Tool callingReasoningImage input

Hugging Face

Large coding-reasoning model for agentic software tasks and RL search

Context: 262K
Released: 3mo ago
Parameters: 35B

Sources:models.dev

Benchmarks

DeepSWE is the primary coding benchmark where an exact model evaluation exists. Other sourced evaluations appear below; missing scores are not zero.

Not evaluated in DeepSWE 1.1 as of 2026-09-22. Results are only shown for exact model versions in the official dataset. Official leaderboard.

No benchmark results have been matched to this model yet.

Specifications

Developer: DeepReinforce
Model family: —
Type: Language model
Predecessor: —
Successor: —
Released: June 25, 2026
Model ID: deepreinforce/ornith-1.0-35b
Openness: Open weights
License: MIT
Parameters: 35B
Input modalities: text, image
Output modalities: text
Modality: text+image->text
Hugging Face: —
Source freshness: Metadata synced Oct 8, 2026, 10:17 PM UTC

Additional source evidence

Source model: deepreinforce/ornith-1.0-35b
Source family: ornith
Last updated: 2026-06-25
Model context: 262K
Open weights: Yes
File attachments: Yes
Reasoning: Yes
Tool calling: Yes
Temperature control: Yes

Downloadable weights

Hugging Face

Published benchmarks

Benchmark	Score	Evaluation details	Evidence
SWE-Bench Verified	75.6 · percent resolved	—	Source
SWE-Bench Pro	50.4 · percent resolved	—	Source
SWE-Bench Multilingual	69.3 · percent resolved	—	Source
Terminal-Bench 2.1	64.2 · percent	Terminus-2	Source
Terminal-Bench 2.1	62.8 · percent	Claude Code	Source
NL2Repo	34.6 · percent	—	Source
Claw-eval	69.8 · percent	—	Source

Official references

Announcement

Complete source record (JSON)

{
  "id": "deepreinforce/ornith-1.0-35b",
  "name": "Ornith 1.0 35B",
  "description": "Large coding-reasoning model for agentic software tasks and RL search",
  "family": "ornith",
  "last_updated": "2026-06-25",
  "attachment": true,
  "reasoning": true,
  "tool_call": true,
  "temperature": true,
  "modalities": {
    "input": [
      "text",
      "image"
    ],
    "output": [
      "text"
    ]
  },
  "limit": {
    "context": 262144
  },
  "benchmarks": [
    {
      "name": "SWE-Bench Verified",
      "score": 75.6,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent resolved"
    },
    {
      "name": "SWE-Bench Pro",
      "score": 50.4,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent resolved"
    },
    {
      "name": "SWE-Bench Multilingual",
      "score": 69.3,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent resolved"
    },
    {
      "name": "Terminal-Bench 2.1",
      "score": 64.2,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent",
      "variant": "Terminus-2"
    },
    {
      "name": "Terminal-Bench 2.1",
      "score": 62.8,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent",
      "variant": "Claude Code"
    },
    {
      "name": "NL2Repo",
      "score": 34.6,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent"
    },
    {
      "name": "Claw-eval",
      "score": 69.8,
      "source": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "metric": "percent"
    }
  ],
  "release_date": "2026-06-25",
  "open_weights": true,
  "license": "MIT",
  "links": [
    {
      "url": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B",
      "label": "Model card",
      "type": "model_card"
    },
    {
      "url": "https://deep-reinforce.com/ornith_1_0.html",
      "label": "Announcement",
      "type": "announcement"
    }
  ],
  "weights": [
    {
      "label": "Hugging Face",
      "url": "https://huggingface.co/deepreinforce-ai/Ornith-1.0-35B"
    }
  ]
}

Infrastructure

Provider availability

All providers

Provider endpoint data will appear after this model's first OpenRouter endpoint sync.

Deployment graph

Can I run this?

Vertical Tabs only shows compatibility when it has structured evidence tying this exact model to hardware, runtime, quantization, memory use, and observed performance.

Browse hardware