Compare/Agents
Compare AI agents
Pick up to 4 agents. Capability benchmarks, real-world OpenRouter usage, base models, license, and pricing render side-by-side.
Looking to compare base models instead? Switch to model compare →
| Hermes Agent Nous Research | Devin Cognition | Codex OpenAI | Claude Code Anthropic | |
|---|---|---|---|---|
| Overview | ||||
| Vendor | Nous Research | Cognition | OpenAI | Anthropic |
| Category | general | coding | coding | coding |
| Release date | 2026-02-01 | — | — | — |
| Source | open source | proprietary | proprietary | proprietary |
| License | Apache-2.0 | — | — | — |
| Pricing | Free OSS | $500/mo Team | Included with ChatGPT Plus/Pro | Included with Claude Pro/Max/Team subscriptions |
| Base models | — | |||
| Capabilities | persistent_memoryself_improvingmulti_llmbrowser+2 | autonomouscloud_runtimemulti_sessionbrowser_use | terminalfile_ioshell_execcloud_runtime | terminalfile_iomulti_file_editshell_exec+1 |
| Capability | ||||
| gpt.buzz Index | 15.0 | 12.4 | 70.0 | 95.4 |
| SWE-bench Verified | — | 35.0% | 85.0% | 80.9% |
| SWE-bench Pro | — | — | 56.8% | 55.4% |
| Terminal-Bench 2.0 | — | — | 77.3% | 65.4% |
| GAIA | 56.0% | — | — | 74.6% |
| WebArena | — | — | — | 64.5% |
| Adoption | ||||
| OpenRouter · daily | 644B | — | — | 73.5B |
| OpenRouter · cumulative | 7.72T | — | — | 3.52T |
| GitHub stars | — | — | — | — |
| Links | ||||
| Links | ||||
Add an agent
Replit AgentReplit · codingmaxHermes AgentNous Research · general✓Claude CodeAnthropic · coding✓CodexOpenAI · coding✓CursorAnysphere · codingmaxGitHub CopilotGitHub · codingmaxClineCline · codingmaxOpenCodeOpenCode · codingmaxDevinCognition · coding✓ManusMonica · generalmaxComputer UseAnthropic · browsermaxOperatorOpenAI · browsermaxCometPerplexity · browsermaxv0Vercel · codingmaxBolt.newStackBlitz · codingmaxLovableLovable · codingmaxOpenClawErik Steinberger · generalmax
Max 4 agents. Remove one to add another.