Models across text & image

Every model. Full transparency.

AgentWorks gives you access to text and image models through a single platform. We publish our exact pricing and margin so you always know what you pay.

Prompt caching saves up to 80%

AgentWorks uses prompt caching where available. Repeated context (system prompts, knowledge base chunks) is cached and re-used - reducing your effective cost by up to 80% on cache hits.

Transparent margin

Our platform margin covers infrastructure, caching layer, gateway reliability, compliance tooling, and support. We keep it transparent so you can compare directly against going to providers yourself.

Free plan

Free accounts have access to Gemini, NanoBanana SLM, and OpenAI models. Upgrade to Enterprise to unlock all models including Mistral, Claude, and premium image models.

Text
ModelProviderContextProvider list priceAgentWorks priceMarginAccess
GPT-4o

Best general-purpose model from OpenAI. Supports vision and function calling.

OpenAI128K€11.20€9.60~15%Free
GPT-4.1

Latest OpenAI flagship optimized for agentic workflows.

OpenAI128K€11.20€9.60~15%Free
Claude 3.5 Sonnet

Best coding and reasoning model. 200K context window.

Anthropic200K€9.80€8.40~14%Enterprise
Gemini 2.0 Flash

Ultra-fast and cost-efficient with 1M token context window.

Google1M€0.84€0.72~14%Free
Gemini 1.5 Pro

Largest context window available - ideal for processing entire documents.

Google2M€7.25€6.20~14%Free
Mistral Large

Top-tier European model. Strong multilingual and reasoning capabilities.

Mistral AI128K€6.75€5.80~14%Enterprise
Mistral Nemo

Fast and efficient European SLM for high-volume tasks.

Mistral AI128K€0.42€0.36~14%Enterprise
Llama 3.3 70B

Best open-weights model at scale. Also available for self-hosted Enterprise.

Meta (hosted)128K€2.45€2.10~14%Enterprise
NanoBanana SLM

AgentWorks-hosted small language model optimized for speed and cost. Great for high-frequency classification and routing tasks.

AgentWorks32K€1.10€0.95~14%Free
MultimodalFunction callingVisionReasoningLong contextAgenticFast1M contextMultilingualEuropeanEfficientOpen weightsInstruction-tunedUltra-fastSLMLow latency
Image
ModelProviderProvider list priceAgentWorks priceMarginAccess
DALL-E 3

High-quality image generation with precise prompt adherence.

OpenAIPer imagePer image~15%Enterprise
Imagen 3

Google's best image model with exceptional photorealism and text rendering.

GooglePer imagePer image~14%Enterprise
Image generationPrompt-basedHDPhotorealisticText rendering

Local LLMs & SLMs (Enterprise)

Enterprise customers can connect their own locally hosted language models or SLMs (e.g. LLaMA, Mistral, custom fine-tunes) running inside their private cloud or on-prem infrastructure. We also offer integrations with our partner slm-works.ai. AgentWorks acts as the orchestration layer - you bring the compute, we provide the governance, pipelines, and tooling.

Talk to our team about local models