⚡ Zen LM
Models

zen4-pro

High capability MoE model. 80B (3B active) with 131K context.

zen4-pro

High Capability

Efficient MoE model for demanding workloads. 80B total parameters with only 3B active per token, delivering strong reasoning at production-grade cost efficiency.

Specifications

PropertyValue
Model IDzen4-pro
Parameters80B (3B active)
ArchitectureMoE
Context Window131K tokens
Tierultra
Input Price$2.70 / 1M tokens
Output Price$2.70 / 1M tokens

Capabilities

  • High-quality reasoning with MoE efficiency
  • Strong performance on demanding benchmarks
  • Balanced cost-to-quality for production workloads
  • 131K context window for extended conversations
  • Hybrid Gated DeltaNet + Gated Attention architecture

API Usage

curl https://api.hanzo.ai/v1/chat/completions \
  -H "Authorization: Bearer $HANZO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "zen4-pro",
    "messages": [{"role": "user", "content": "Analyze the trade-offs between microservices and monoliths."}]
  }'
from hanzoai import Hanzo

client = Hanzo(api_key="hk-your-api-key")

response = client.chat.completions.create(
    model="zen4-pro",
    messages=[{"role": "user", "content": "Analyze the trade-offs between microservices and monoliths."}],
)
print(response.choices[0].message.content)

See Also

  • zen4-thinking -- Same architecture with chain-of-thought reasoning
  • zen4-max -- Larger MoE for frontier reasoning
  • Pricing -- Full pricing table

On this page