Build coding & tool-use agents

Understand — Gemini 3 Flash; Reason — DeepSeek-V4-Pro; Execute — Claude Fable 5. Each row above expands with the model's live pricing, capability chips, and sample outputs.

DeepSeek-V4-Pro runs on Runware's own optimized compute (the platform's open-weight tier, billed on compute time) — check each model page for weight availability and license terms before self-hosting. The remaining picks are partner-served.

The lowest listed starting price in this stack is Gemini 3 Flash from $0.5 / 1M per Input tokens (text, image, video); prices are read live from the catalog and scale with resolution, duration, and quality tier. The Production notes below cover how pricing is measured.

Published input rates run from $0.127 per 1M tokens (MiniMax M2.5) to $10 per 1M (Claude Fable 5). Route the bulk of agent traffic to the cost-efficient tier and escalate to the frontier tier only for the steps that fail there.

DeepSeek-V4-Pro and Claude Fable 5 both offer 1M-token contexts, and DeepSeek's cache reads drop to $0.079 per 1M tokens. Long-running agents should be architected around cache reuse rather than resending full history at input rates.

Every price on this page is the model's published rate from the live Runware catalog, using the cheapest listed configuration unless stated otherwise. Prices vary with resolution, duration, quality tier, or token volume, so check the pricing table on each model page before estimating unit economics.

The Runware catalog does not publish per-model latency figures, so this page does not quote end-to-end timings. Where a model's own description commits to speed (for example sub-second generation or realtime streaming), that claim is repeated here. For anything else, benchmark the exact models in the Playground with your own payload sizes before committing to an SLA.

Models are addressed by versioned AIR identifiers, so a workflow pinned to specific model versions keeps producing the same behaviour as new versions ship. Adopt upgrades deliberately by re-running your evaluation set against the new version before switching production traffic.

A blueprint bundles the small set of models you would actually wire together for one production use case, instead of ranking a whole category. Each pick covers one stage of the workflow and links to its model page for full schema, pricing, and examples.

Picks are live models from the Best LLMs (25 models) and Best Coding Agents (10) collections, chosen to cover four price and capability tiers: cost-efficient agentic coding, multimodal input, 1M-token context with caching, and frontier long-horizon capability.