Category
Agentic Coding Model Models
The models we recommend for long-running, agentic coding tasks — multi-step planning, tool use, and code edits.
3 models
← All categories
Top Agentic Coding Model models
| Model | Pricing |
|---|---|
| Claude Sonnet 4.6 Anthropic | $3.00/1M input |
| DeepSeek V4 Pro preview DeepSeek | $0.44/1M input |
| GLM 5.2 Z Ai | $1.40/1M input |
How these models were selected
Picks prioritize multi-step coding, tool invocation, repository navigation, and the ability to recover from intermediate failures during autonomous work.
When to use a different shortlist
Agentic reliability depends on the harness as much as the model. Permission controls, sandboxing, context management, and retry policy can change outcomes dramatically.
Open each model page to verify current pricing, context limits, source links, and known limitations before choosing a provider.