Apodex 1.1
Apodex 1.1 by Apodex is an agentic model for long-horizon work, using Environment Scaling, Agentic Coordination Scaling, and AgentOS state tracking.
Model Overview
Capabilities, design details, and architectural traits
Apodex 1.1 - A Heavy-Duty Solver for Long-Running, Verifiable Work
Apodex 1.1 is a general-purpose model and execution system built around what its creators call working capability: sustained, verifiable progress toward a real-world objective, measured in completed work rather than isolated responses. It is designed for tasks that unfold over long horizons, where the model must operate on files, run and debug code, revise plans as observations change, recover from failure, and deliver artifacts someone else can inspect and continue using.
The model develops this capability along two documented scaling dimensions:
- Environment Scaling expands the diversity, fidelity, and verifiability of the executable file, search, and code worlds the model acts in. Actions, file changes, code outputs, recovery decisions, and artifacts become part of the learning distribution rather than textual descriptions.
- Agentic Coordination Scaling trains agents to decompose long-horizon objectives, delegate parallel work, integrate asynchronous results, and reorganize unfinished branches.
What Sets Apodex 1.1 Apart
| Differentiator | Description |
|---|---|
| Working capability | Success is defined as sustained, verifiable progress toward an objective, with failure recovery and a delivery contract, not single-turn answer quality. |
| AgentOS runtime | A shared execution harness that maintains task state and provenance across tools and agents, and lets a user step in mid-task to redirect the work. |
| Agent Team coordination | Adaptive parallel effort across evidence gathering, file analysis, implementation, verification, and counteranalysis. |
| Two-axis training | Environment trajectories and coordination traces are both turned into reliable behavior through training. |
| Apodex 1.1 Mini | A 35B-parameter open-weight sibling that retains strong working capability in a locally deployable form. |
Target Domains
Apodex 1.1 is aimed at complex professional work, finance, scientific research, mathematics, coding, and search. It accepts text, images, PDFs, and CSV/Excel input with a 256K-token context window. The long-term goal behind the system is a Heavy-Duty Solver that can take responsibility for ambitious, long-running, verifiable work.
Benchmark Performance
Independent evaluations · Artificial Analysis
Accuracy & Capability Details
Compare Models Side-by-Side
Evaluate specifications, pricing, and independent benchmark indices
| Model Details | |||
|---|---|---|---|
| General Info | |||
| Provider | Apodex | Anthropic | Anthropic |
| Release Date | August 30, 2026 | September 1, 2026 | July 24, 2026 |
| Knowledge Cutoff | - | - | May 2026 |
| Context & Limits | |||
| Context Window | 262K | - | 1M Best Context Window |
| Pricing (per 1M tokens) | |||
| Input Pricing | $0.30 Best Input Pricing | $10 | $5 |
| Output Pricing | $3 Best Output Pricing | $50 | $25 |
| Modalities | |||
| Inputs | text | textimagefile | textimage |
| Outputs | text | text | text |
| Benchmarks (0-100) | |||
| Intelligence Index | 44.0 | 65.7 Best Intelligence Index | 63.1 |
| Coding Index | 60.8 | 81.6 Best Coding Index | 78.0 |
| Agentic Index | 36.6 | 61.3 Best Agentic Index | 59.2 |
GPQA Benchmark
Graduate-level reasoning and expert Q&A evaluation.
Humanity's Last Exam
Extremely difficult logical reasoning and knowledge.
Long Context Reasoning
Logical reasoning over long context windows.
Independent evaluation data provided by Artificial Analysis. To view the latest benchmarks and full details, visit their official site.