Instructions to use solarpush/push-26b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use solarpush/push-26b with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf solarpush/push-26b
# Run inference directly in the terminal:
llama cli -hf solarpush/push-26b

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf solarpush/push-26b
# Run inference directly in the terminal:
llama cli -hf solarpush/push-26b

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf solarpush/push-26b
# Run inference directly in the terminal:
./llama-cli -hf solarpush/push-26b

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf solarpush/push-26b
# Run inference directly in the terminal:
./build/bin/llama-cli -hf solarpush/push-26b

Use Docker

docker model run hf.co/solarpush/push-26b

LM Studio
Jan
Ollama
How to use solarpush/push-26b with Ollama:
```
ollama run hf.co/solarpush/push-26b
```
Unsloth Desktop

How to use solarpush/push-26b with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf solarpush/push-26b

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "solarpush/push-26b"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use solarpush/push-26b with Docker Model Runner:
```
docker model run hf.co/solarpush/push-26b
```

Lemonade

How to use solarpush/push-26b with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull solarpush/push-26b

Run and chat with the model

lemonade run user.push-26b-{{QUANT_TAG}}

List all available models

lemonade list

Hermes Agent

How to use solarpush/push-26b with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf solarpush/push-26b

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default solarpush/push-26b

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use solarpush/push-26b with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf solarpush/push-26b

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "solarpush/push-26b" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

push:26b — Calibrated System-1 Decision Model (26B-A4B MoE GGUF)

push:26b (solarpush/push-26b) is a high-precision, calibrated System-1 Decision Model (llm-logits-v1) designed for single-forward-pass structured decision-making across three universal primitives:

choice — Multi-way categorical selection (A..Z, 2 to 26 options),
noul — Binary epistemic / policy / safety / entailment verification (Yes vs No),
score — Ordinal rubric & severity evaluation (0..9, 3- to 5-level calibrated scales).

Built on a sparse 26B total / 4B active parameter Mixture-of-Experts (MoE) GGUF backbone (UD-Q3_K_M, 12.42 GiB on disk, ~15.0 GB VRAM) with an untied, subspace-orthogonal Q6_K (ttype=14) decision head and cardinality-conditioned temperature scaling (temperature_by_options), push:26b delivers frontier-grade accuracy and calibration in ~223 ms p50 on a single 16 GB consumer GPU.

1. Official Ollaya Decision Benchmark (`bench.ollaya.dev` — 14 Tasks, 5,880 Questions)

Evaluated strictly zero-shot / out-of-domain across all 14 official bench.ollaya.dev tasks (3,880 Public benchmark questions across 13 datasets + 2,000 typed-decisions workflow questions across 400 multi-question JSON states = 5,880 total questions, zero benchmark train/dev/test overlap):

Model	Overall Skill `[95% CI]`	Public Skill (13 Tasks)	Public Macro Acc	Typed-Decisions Acc (Skill)	Public Pooled ECE (lower is better)	Typed ECE (lower is better)	Global Brier	Latency (p50)	VRAM	License
`push:26b` (26B-A4B MoE)	`62.7` `[60.8, 64.7]`	`63.1`	`75.4%`	`71.4%` (`58.1`)	`2.8%` (`0.028`)	`2.0%` (`0.020`)	`0.368`	`223 ms`	`15.0 GB`	`Apache-2.0`
`winnow:e4b` (Reference 4B)	`59.8` `[58.0, 61.5]`	`58.9`	`72.7%`	`72.4%` (`59.5`)	`3.4%` (`0.034`)	`3.2%` (`0.032`)	`0.380`	`46 ms`	`6.5 GB`	`Apache-2.0`
`push:3b` (Adapted 3B)	`46.0` `[43.9, 48.2]`	`47.1`	`64.8%`	`53.9%` (`32.5`)	`3.1%` (`0.031`)	`15.3%` (`0.153`)	`0.520`	`86 ms`	`4.2 GB`	`Apache-2.0`

Accuracy by Decision Primitive (5,880 Questions)

Choice (Multi-Option Selection): 76.8%
Noul (Binary Verification & Guardrails): 82.2%
Score (Ordinal Rubric Evaluation): 60.5%

2. Per-Task Breakdown (14 Tasks — 5,880 Questions)

Benchmark Task	Primitive	Questions	`push:26b` Accuracy	`push:26b` Skill (0–100)	`winnow:e4b` Accuracy (Skill)	Delta vs `winnow:e4b`	`push:26b` ECE (10-bin)	`push:26b` Brier	Median Latency
`squad2` (Unanswerable QA)	`noul`	299	`87.3%` (up to `87.6%`)	`74.6`	`79.6%` (`59.2`)	`+7.7%` (`+15.4` Skill)	`0.064`	`0.211`	`122 ms`
`massive-en-US` (18-Way Intent)	`choice`	350	`86.9%`	`86.1`	`84.6%` (`83.7`)	`+2.3%` (`+2.4` Skill)	`0.047`	`0.222`	`124 ms`
`summeval-consistency` (Factual Consistency)	`score`	144	`86.1%`	`82.6`	`83.3%` (`79.2`)	`+2.8%` (`+3.4` Skill)	`0.187`	`0.274`	`159 ms`
`massive-de-DE` (18-Way Intent DE)	`choice`	350	`85.4%`	`84.6`	`83.7%` (`82.8`)	`+1.7%` (`+1.8` Skill)	`0.071`	`0.246`	`125 ms`
`boolq` (Reading Verification)	`noul`	300	`84.0%` (up to `85.3%`)	`68.0`	`86.0%` (`72.0`)	`-2.0%`	`0.096`	`0.228`	`107 ms`
`civil_comments` (Toxicity Detection)	`noul`	300	`82.0%` (up to `82.3%`)	`64.0`	`75.7%` (`51.3`)	`+6.3%` (`+12.7` Skill)	`0.119`	`0.268`	`95 ms`
`paws` (Adversarial Paraphrase)	`noul`	250	`80.4%`	`60.8`	`84.0%` (`68.0`)	`-3.6%`	`0.058`	`0.300`	`99 ms`
`multinli` (3-Way NLI)	`choice`	299	`79.6%` (up to `81.3%`)	`69.4`	`82.9%` (`74.4`)	`-3.3%`	`0.044`	`0.320`	`92 ms`
`aegis2` (AI Safety Guardrail)	`noul`	250	`78.0%` (up to `79.2%`)	`56.0`	`80.8%` (`61.6`)	`-2.8%`	`0.094`	`0.324`	`101 ms`
`pubmedqa` (Biomedical 3-Way QA)	`choice`	250	`74.0%` (up to `74.4%`)	`61.0`	`69.6%` (`54.4`)	`+4.4%` (`+6.6` Skill)	`0.076`	`0.359`	`140 ms`
`vitaminc-dev` (3-Way Fact Verification)	`choice`	599	`72.5%`	`58.7`	`69.3%` (`53.9`)	`+3.2%` (`+4.8` Skill)	`0.062`	`0.414`	`102 ms`
`typed-decisions` (400 JSON Workflows)	`mixed`	2000	`71.4%` (up to `71.8%`)	`58.1`	`72.4%` (`59.5`)	`-1.0%`	`0.020`	`0.402`	`294 ms`
`summeval-relevance` (5-Level Relevance)	`score`	240	`48.3%`	`35.4`	`44.6%` (`30.7`)	`+3.7%` (`+4.7` Skill)	`0.046`	`0.659`	`189 ms`
`helpsteer2` (5-Level Helpfulness)	`score`	249	`40.2%` (up to `41.4%`)	`25.2`	`37.3%` (`21.7`)	`+2.9%` (`+3.5` Skill)	`0.154`	`0.741`	`167 ms`
OVERALL (14 Tasks)	All	5880	`75.1%` (`75.4%` Public)	`62.7`	`72.6%` (`59.8`)	`+2.7%` Public (`+2.9` Skill)	`0.083` (`0.028` Pooled)	`0.368`	`223 ms`

`typed-decisions` Enterprise Workflow Breakdown (2,000 Questions / 400 JSON States)

customer_service: 77.6%
invoice_processing: 74.6% (up to 76.0%)
security_incidents: 72.6%
agent_trace_observability: 61.6% (up to 64.2%)
Calibration: ECE = 0.020 (2.0%) | Soft-CE = 1.010

3. Training & Calibration Methodology (100% Zero-Benchmark-Leakage)

push:26b was adapted with Trainaya using strictly generalist, out-of-benchmark decision datasets (ZefanCai/Open-Jev-v1.1, SargeDev/jev-distill-corpus-v3, tasksource/tasksource-jev-typed-decisions with all 14 benchmark dataset sources strictly excluded, plus balanced cyclic $3 \times 3$ ABC counterfactual triplets):

Untied Q6_K (ttype=14) Output Head Materialization:
- Unties output.weight ([262144, 2816]) from token_embd.weight so the 62 decision rows (A..Z, space-prefixed ▁A..▁Z, and 0..9) are updated in high-precision Q6_K without perturbing backbone token embeddings.
Biorthogonal Dual Subspace Projection (U, U_s, vt_bin, vt_score_prot):
- Solves separate orthogonal projections for 3-way NLI, 3-way Fact Verification, 3-way QA, 4- and 5-option workflow choices, binary noul verification, and 5-level ordinal rubrics so updates to multi-option choice produce zero drift (0.000000) on protected binary noul and ordinal score subspaces.
Fisher-ARD Kernel Ridge + Composite Soft-Target Distillation (KL + Brier + RPS + Ordinal + Softcap):
- Blends hard gold labels with soft teacher distributions (0.65 * y_hard + 0.35 * y_soft) across multi-scale ridge ensembles (λ ∈ {20, 40, 75}) followed by subspace-protected composite KL + Brier + Ranked Probability Score (RPS) polish respecting the architecture's native final_logit_softcap = 30.0.
Cardinality-Conditioned Temperature Calibration (temperature_by_options):
- Calibrates distinct post-softcap temperatures per primitive and option count (noul:2 = 8.91, choice:2 = 6.35, choice:3-5 = 8.2485, choice:6-10 = 6.0325, choice:11+ = 4.6355, score:3-5 = 10.3), achieving a Pooled Public ECE of 2.8% (0.028).

4. Repository Contents

File	Size	Description
`model.gguf`	`12.42 GiB`	26B-A4B MoE GGUF (`UD-Q3_K_M` backbone + untied `Q6_K` calibrated `output.weight` decision head)
`decision.json`	`2.5 KB`	Ollaya `llm-logits-v1` decision specification (`labels`, `template`, `temperature_by_options`, `llama` runtime config)
`calibration.json`	`751 B`	Post-hoc temperature scaling parameters and out-of-fold / post-quantization accuracy metrics
`ollaya_config.json`	`302 B`	Ollaya runtime manifest configuration
`Modelfile`	`559 B`	Deterministic single-token Ollama `Modelfile` (`temperature 0`, `top_k 1`, `num_predict 1`)

5. Quickstart (Ollaya & Ollama)

Option A: Register in both Ollaya and Ollama via `trainaya`

git clone https://huggingface.co/solarpush/push-26b
cd push-26b
trainaya register --model . --name push:26b --runtime both

Option B: Query via Ollaya `/v1/systemone` API

curl -s http://127.0.0.1:11435/v1/systemone \
  -H "Content-Type: application/json" \
  -d '{
    "model": "push:26b",
    "state": "Context: Aspirin inhibits platelet aggregation via irreversible COX-1 acetylation. In a randomized trial of 1,200 patients, low-dose aspirin reduced recurrent thrombotic events by 24% (p = 0.003).",
    "questions": [
      {
        "id": "q1",
        "type": "noul",
        "question": "Is the reduction in recurrent thrombotic events statistically significant at p < 0.01?",
        "options": {"true": "Yes, statistically significant", "false": "No, not significant"}
      }
    ]
  }'

Option C: Run with Ollama

ollama create push:26b -f Modelfile
ollama run push:26b

Downloads last month: 477

GGUF

Model size

26B params

Architecture

gemma4

Hardware compatibility

We're not able to determine the quantization variants.

View all variants

Model tree for solarpush/push-26b

Base model

google/gemma-4-26B-A4B

Finetuned

google/gemma-4-26B-A4B-it

Quantized

(412)

this model

solarpush
/

push-26b

push:26b — Calibrated System-1 Decision Model (26B-A4B MoE GGUF)

1. Official Ollaya Decision Benchmark (`bench.ollaya.dev` — 14 Tasks, 5,880 Questions)

Accuracy by Decision Primitive (5,880 Questions)

2. Per-Task Breakdown (14 Tasks — 5,880 Questions)

`typed-decisions` Enterprise Workflow Breakdown (2,000 Questions / 400 JSON States)

3. Training & Calibration Methodology (100% Zero-Benchmark-Leakage)

4. Repository Contents

5. Quickstart (Ollaya & Ollama)

Option A: Register in both Ollaya and Ollama via `trainaya`

Option B: Query via Ollaya `/v1/systemone` API

Option C: Run with Ollama

Model tree for solarpush/push-26b

Datasets used to train solarpush/push-26b

push:26b — Calibrated System-1 Decision Model (26B-A4B MoE GGUF)

1. Official Ollaya Decision Benchmark (bench.ollaya.dev — 14 Tasks, 5,880 Questions)

Accuracy by Decision Primitive (5,880 Questions)

2. Per-Task Breakdown (14 Tasks — 5,880 Questions)

typed-decisions Enterprise Workflow Breakdown (2,000 Questions / 400 JSON States)

3. Training & Calibration Methodology (100% Zero-Benchmark-Leakage)

4. Repository Contents

5. Quickstart (Ollaya & Ollama)

Option A: Register in both Ollaya and Ollama via trainaya

Option B: Query via Ollaya /v1/systemone API

Option C: Run with Ollama

Model tree for solarpush/push-26b

Datasets used to train solarpush/push-26b

1. Official Ollaya Decision Benchmark (`bench.ollaya.dev` — 14 Tasks, 5,880 Questions)

`typed-decisions` Enterprise Workflow Breakdown (2,000 Questions / 400 JSON States)

Option A: Register in both Ollaya and Ollama via `trainaya`

Option B: Query via Ollaya `/v1/systemone` API