Devin
Researched 1d agoPlatformTier 2Cognition
Devin offers 2 tracked models (0 with output token pricing). This catalog covers coding, agents, and json / tool use; open any model detail page for benchmarks, batch tiers, and migration prompts.
Covers 3 workload areas across 2 tracked models; last verified 2026-09-10.
Use it for
- Operators routing coding, agents, and json / tool use workloads through this API
Do not use it for
- Final benchmark picks without opening the relevant model detail page
- Strict price-per-token comparisons until output pricing is sourced
Tracked models
2
Models available through this provider
Priced output routes
0
Output pricing not yet tracked
Cheapest output
Unknown
Output pricing not yet tracked
Batch-ready models
0
No batch pricing tracked
Latest model release
2026-09-10
1d since newest release
Freshness
2026-09-10
Researched 1d ago
Information
Devin is Cognition's autonomous AI software engineer platform, providing Web, Desktop, CLI, and API access to Cognition-trained SWE models for long-horizon coding and agentic software-engineering work.
Where this host wins
- Coding: 2 tracked models with SWE-bench / HumanEval-style scores.
- Agentic: 2 tracked models with BFCL, tau-bench, and SWE-bench tool-use coverage.
- JSON/tool-use: 2 tracked models with BFCL / Nexus strict-JSON routing coverage.
Getting started
Platform Overview
Devin is a hosted coding-agent platform rather than a token-metered model API. Users sign up at app.devin.ai, install the Devin CLI or Desktop client, or create programmatic sessions through the Devin REST API. Model selection resolves family aliases such as swe to the latest SWE version; Devin pricing is plan, quota, and on-demand credit based rather than published per-model token pricing.