orcarouter/OrcaSAQ-2-27B — new model trending #30 on Hugging Face

<p align="center">The point is not 3-bit.</p>

<p align="center"><strong>The point is what survives at 3-bit.</strong></p>

---

## BF16 fidelity

All numbers below are measured using **these exact OrcaSAQ2 weights** against the BF16 reference through the same evaluation path.

### WikiText-2

**16,376 predicted tokens**

| Build | Size | Decoder Bits | Mean KLD ↓ | Top-1 Agreement ↑ | PPL ↓ |

|---|---:|---:|---:|---:|---:|

| Qwen3.8-27B BF16 | 54 GB | 16 | — | 100% | 5.6468 |

| **OrcaSAQ2 27B** | **12.3 GB** | **3.21** | **0.031** | **93.2%** | **5.6482** |

### Perplexity

```text

BF16 5.6468 ████████████████████████████████████████

OrcaSAQ2 5.6482 ████████████████████████████████████████

Delta: +0.02%

```

### Top-1 agreement

```text

OrcaSAQ2 vs BF16

███████████████████████████████████████████████░░░ 93.2%

```

### Model footprint

```text

Qwen3.8-27B BF16

██████████████████████████████████████████████████ 54.0 GB

OrcaSAQ2

███████████ 12.3 GB

```

**77.2% smaller.**

---

## Long-horizon agents

### Low-bit fidelity matters more as the horizon grows.

Short benchmarks can hide small degradation.

Agents cannot.

A small model error can change a tool call.

That changes the environment state.

The changed state affects every decision that follows.

```text

Plan

↓

Act

↓

Observe

↓

Decide

↓

Recover

↓

Repeat

↓

...

↓

Task Success

```

Across long trajectories, small errors can compound into large behavioral differences.

That makes long-horizon execution an especially useful stress test for compressed reasoning models.

OrcaSAQ2 performs **strongly on long-horizon workloads relative to models in its deployment and parameter class**, despite operating from a **12.3 GB checkpoint**.

This makes it particularly suitable for:

- coding agents

- terminal agents

- browser agents

- computer-use agents

- security agents

- repository-scale tasks

- multi-tool workflows

- failure recovery

- long-running stateful execution

### Why this matters

Perplexity asks:

> **How similar is the next-token distribution?**

Long-horizon evaluation asks:

> **Can the model still finish the job after many decisions?**

Text extracted automatically; images, tables and formatting may be missing. Original: https://huggingface.co/orcarouter/OrcaSAQ-2-27B