Best Coding Agents
Models optimized for coding, agentic tool use, and multi-step workflows. Ideal for writing, debugging, and refactoring code, executing tool calls, and orchestrating complex long-running automated pipelines.
The picks.Reliable models for coding, tool use, and agentic workflows
Claude Fable 5
Frontier multimodal agentic model for long-horizon coding, research, and complex professional work
GLM-5.3
Flagship coding and agentic LLM with stronger long-horizon execution and security-oriented reasoning
Kimi K3
Flagship multimodal reasoning LLM for long-horizon coding, knowledge work, and tool-rich agent workflows
GPT-5.6 Sol
Frontier reasoning LLM for complex professional work, long context, and full tool-using workflows
GLM-5.2
Flagship long-horizon coding LLM with usable 1M context and stronger end-to-end engineering execution
Claude Opus 4.7
Frontier multimodal reasoning model for advanced coding, long-running agents, and complex knowledge work
Claude Opus 4.8
Frontier multimodal reasoning model for advanced coding, agent systems, and long-context knowledge work
Claude Sonnet 4.6
Versatile multimodal language model for coding, agents, computer use, and production reasoning
GPT-5.6 Terra
Balanced GPT-5.6 model for professional workloads that need strong reasoning with lower cost
Kimi K2.6
Open frontier multimodal LLM for coding, long-horizon execution, and tool-rich workflows
GLM-5.1
Flagship agentic coding model with 200K context, deep thinking, and long-horizon task execution
Gemini 3.6 Flash
More efficient multimodal Flash model for coding, knowledge work, and agentic execution
GPT-5.6 Luna
Cost-efficient GPT-5.6 variant for fast, high-volume text and vision workflows
DeepSeek-V4-Pro-0423
High-capability frontier LLM with 1M context, stronger agent performance, and dual thinking modes
GPT-5.5
Frontier reasoning LLM for complex coding, long-context work, and tool-using professional tasks
Gemini 3.5 Flash
Frontier multimodal reasoning model for agentic and coding workflows
Gemini 3.5 Flash-Lite
Fast high-throughput Gemini model for low-latency agentic and document-heavy workloads
GPT-5.4
Flagship reasoning LLM with 1M context, native computer use, and high factual accuracy
GPT-5.4 Mini
Efficient reasoning LLM with 400K context for coding assistants and subagent workflows
Claude Haiku 4.5
Fast, cost-efficient multimodal language model for low-latency agents and scaled reasoning workloads
GPT-5.4 Nano
Ultra-low-latency LLM for high-volume classification, extraction, and lightweight automation
Open weights.Hosted alternatives for this stack
Open-weight models covering the same tasks as this stack, running on Runware's own optimized compute and billed on compute time. License terms vary per model — check each model page before self-hosting.
GLM-5.3-Flash
Efficient multimodal GLM for coding, agents, and long-context visual understanding
Gemma 4 31B
Open 31B multimodal reasoning model for coding, long context, and agentic workflows
DeepSeek-V4-Flash-0731
Fast frontier LLM with 1M context, tool use, and dual thinking modes
Common questions
Yes — GLM-5.3, Kimi K2.6, GLM-5.1, DeepSeek-V4-Pro-0423 run on Runware's own optimized compute (the platform's open-weight tier, billed on compute time). Check each model page for license terms before self-hosting.
GLM-5.3, released August 2026 per the live catalog. Membership updates automatically as the catalog publishes new models to this collection.
Production notes
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.
About this collection
Models optimized for coding, agentic tool use, and multi-step workflows. Ideal for writing, debugging, and refactoring code, executing tool calls, and orchestrating complex long-running automated pipelines.
Membership comes directly from the Runware catalog: the 21 models on this page are the live catalog's own membership for the "Best Coding Agents" collection. Names, descriptions, pricing, capability chips, samples, and guides are all read live from the catalog — nothing here is hand-curated.