← Products
Model 01 — Shipped

Arcle V1

Unified omni model.

One model. Seven capabilities. 5.84 billion parameters. Text, images, documents, speech and audio all flow through one nn.Module and one shared representation space — a single forward pass, a single set of weights, a single model file.

Access the Model

Fig. 1 — What it is

Not a pipeline.
One model.

Most systems that advertise "multimodal" are orchestration layers — a vision model here, a speech model there, a language model in the middle, glued together by application code.

Arcle V1 instead projects every modality into the same 2,560-dimensional hidden space that the language core reasons in, so the model genuinely sees an image and hears audio in the same way it reads text. It is not a router, not a pipeline, and not a collection of separate models called one after another behind an API.

Fig. 2 — Specifications

Complete
specifications

SpecificationValue
Total parameters5,840,000,000 (exact)
Language coreHybrid Mamba-2 SSM + attention
Core hidden dimension2,560
Context window2,097,152 tokens (2M)
Image generation512 × 512, 8-step LCM sampling
Image understandingat 384 px
Document / OCR encoder256 tokens per page
Speech recognitionEncoder, 30-second windows
Speech synthesis24 kHz
Languages18+
Precisionbfloat16

Fig. 3 — Every feature

Thirteen
capabilities

3.1

Text generation & conversation

Full instruction-following chat with reasoning, explanation, summarisation, rewriting and open-ended dialogue.

3.2

Mathematical & logical reasoning

Step-by-step working on multi-step arithmetic, algebra and word problems, reinforced through preference optimisation on mathematics-specific data.

3.3

Code generation & understanding

Writes, explains and reasons about code, reinforced with dedicated code preference data.

3.4

Text-to-image generation

Original 512×512 images from a prompt in 8 sampling steps — fast enough for interactive use. Conditioned by the language core itself.

3.5

Image understanding & VQA

Describes images, answers questions about them, and reasons over visual content.

3.6

Document reading & OCR

Reads text out of scanned pages, PDFs, forms, charts and screenshots. 95% content-word recall with no repetition artefacts.

3.7

Speech-to-text (ASR)

Transcribes spoken audio through the Whisper Medium encoder feeding a window-level connector. Measured at 9.7% word error rate.

3.8

Text-to-speech

Generates natural 24 kHz speech through the Voice engine, conditioned by a trained style connector.

3.9

Audio understanding

Spoken input feeds the same reasoning core, so it can be reasoned about, answered and acted on in one pass.

3.10

Long-context processing

A 2,097,152-token window enabled by the hybrid Mamba-2 backbone — sequence length handled natively, not at attention's quadratic cost.

3.11

Multilingual capability

18+ languages, with deliberate Hindi strength.

3.12

Deep India knowledge

Explicitly trained to know India — its geography, culture, history, systems and languages — rather than treating it as a footnote.

3.13

Single-file deployment

Every capability lives in one nn.Module. One model load — no service mesh, no inter-model latency, no orchestration layer.

Fig. 4 — Evaluation

Benchmark
scores

BenchmarkScoreMeasures
GSM8K77.5%multi-step mathematical reasoning
ARC-Easy80.0%grade-school science reasoning
HellaSwag67.0%commonsense sentence completion
ARC-Challenge48.5%hard science reasoning
MMLU43.5%broad multi-domain knowledge
95.0%
OCR — in-distribution
content-word recall
95.4%
OCR — adapter active
content-word recall
90.3%
Speech recognition
9.7% word error rate
89.6%
Audio pathway
understanding
57.0%
Image understanding
caption overlap
5.84B
Parameter count
exact ✓
Standout result

GSM8K at 77.5% is well above what a 3B-class parameter budget would normally suggest — Arcle's own capability from the reasoning-SFT stage, not inherited from base weights.

Fig. 5 — Engineering

How it was
built

Modality-grouped batching

Every batch draws from one modality, turning five tiny sub-batches per step into one full-width forward pass.

Length-bucketed sampling

Near-uniform sequence lengths per batch so padding is not wasted compute, while composition still reshuffles every epoch.

Checkpoint rotation & crash recovery

Training survives interruption and resumes on its own weights.

Exact parameter budgeting

LoRA rank and connector widths solved to land the total on 5,840,000,000 precisely.

Trained over 180 hours on one GPU. It generates images from text. It reads documents. It listens and transcribes. It speaks. It reasons through mathematics. It knows India. And it does all of it in one forward pass through one set of weights.

Free for the people who need it most.

Weights are open on Hugging Face.