Model Overview

LW AI provides access to high-quality AI models through a unified, OpenAI-compatible API. Choose the right model for your task based on context length, capabilities, and cost.

Available Models

ModelContextInput ($/M)Output ($/M)
DeepSeek V4 Pro1.0M--
DeepSeek V4 Flash1.0M--
DeepSeek Reasoner128K--
DeepSeek Chat128K--
GLM-5.21.0M--
GLM-5.1200K--
GLM-5200K--
Yantronic V1 Pro1.0M--
Astra V1 Pro1.0M--

Note: Prices are listed per million tokens. Contact sales for enterprise pricing and volume discounts.

Model Categories

Advanced Reasoning

DeepSeek V4 Pro, DeepSeek Reasoner, and GLM-5.1 are built for complex reasoning, coding, and mathematical tasks with large context windows.

Fast & Cost-Effective

DeepSeek V4 Flash and DeepSeek Chat offer lower latency and cost, ideal for high-volume applications and general-purpose tasks.

Domain-Specific

Yantronic V1 Pro is optimized for finance and insurance reasoning, delivering superior accuracy on domain-specific tasks.

Astra V1 Pro — Fusion & Orchestration

Astra V1 Pro is LW AI's multi-model fusion LLM that uses intelligent routing and ensemble inference to automatically select and combine the best models for each query, optimizing for cost, speed, and accuracy.

Feature Support

ModelStreamingFunction Calling1M Context
DeepSeek V4 Pro
DeepSeek V4 Flash
DeepSeek Reasoner
DeepSeek Chat
GLM-5.1
GLM-5
Yantronic V1 Pro
Astra V1 Pro

Choosing a Model

Your TaskRecommended Model
Complex coding / mathDeepSeek V4 Pro
High-volume, low-latencyDeepSeek V4 Flash
Math proofs / logic chainsDeepSeek Reasoner
General-purpose, cost-effectiveDeepSeek Chat
Agentic engineering / codingGLM-5.1
Systems engineeringGLM-5
Finance / insurance domainYantronic V1 Pro
Automatic model selectionAstra V1 Pro

Rate Limits

Default rate limits vary by model and account tier. See Rate Limits for details.