New

Work in progress: Agents Directory has just launched. Stay tuned, more content is on the way.

Sign In

Rankings

Independent model, benchmark, and agent rankings for AI coding, showing what actually leads right now. Refreshed regularly.

Updated June 2026

Browse modelsBenchmarksProviders

Best models for your agent

Best free models for Hermes

Want to run Hermes without paying per token? Every model here costs $0 on OpenRouter (rate limits apply) and still handles real agent work.

Best models for Claude Code

Claude Code is tuned for the Claude family, but it's worth knowing where each model lands on capability and cost for your workload.

Best models for Codex

Codex only runs OpenAI models, so the real choice is which GPT tier fits your work. Here's how the three options in the picker compare on capability and cost.

Best models for Hermes

Hermes runs your skills locally and leans on the model for planning and skill use. These are the models that pair best with it right now, ranked by real agentic-coding performance and day-to-day reliability.

Best models for OpenClaw

OpenClaw lives in your messaging apps and acts on your behalf, so it rewards models with reliable skill use, long sessions, and sane costs. These picks balance capability against the always-on token bill.

Best open-source models for Hermes

Open-weight models you can inspect, fine-tune, and self-host. Ideal for privacy-sensitive or air-gapped Hermes setups.

Best models for agents

  • 1
    ClaudeClaude Fable 5
    ProviderAnthropic
    Open—
    CursorBench72.9%
    Context1M
    Input price$10/M
  • 2
    ClaudeClaude Opus 4.7
    ProviderAnthropic
    Open—
    CursorBench64.8%
    Context1M
    Input price$5/M
  • 3
    OpenAIGPT-5.5
    ProviderOpenAI
    Open—
    CursorBench64.3%
    Context1.05M
    Input price$5/M
  • 4
    ClaudeClaude Opus 4.8
    ProviderAnthropic
    Open—
    CursorBench63.8%
    Context1M
    Input price$5/M
  • 5
    CursorComposer 2.5
    ProviderCursor
    Open—
    CursorBench63.2%
    Context200K
    Input price$0.5/M
  • 6
    CursorComposer 2
    ProviderCursor
    Open—
    CursorBench52.2%
    Context200K
    Input price$0.5/M
  • 7
    GeminiGemini 3.5 Flash
    ProviderGoogle
    Open—
    CursorBench49.8%
    Context1.049M
    Input price$1.5/M
  • 8
    ClaudeClaude Sonnet 4.6
    ProviderAnthropic
    Open—
    CursorBench49%
    Context1M
    Input price$3/M
  • 9
    MoonshotAIKimi K2.6
    ProviderMoonshot
    Open
    CursorBench47.6%
    Context262K
    Input price$0.68/M
  • 10
    MoonshotAIKimi K2.5
    ProviderMoonshot
    Open
    CursorBench31.9%
    Context262K
    Input price$0.35/M

Benchmark leaderboards

Artificial Analysis Intelligence Index

The most-cited composite intelligence score: a 0–100 index combining knowledge, reasoning, math, coding, and agentic evaluations (GPQA Diamond, HLE, IFBench, SciCode, Terminal-Bench Hard, τ²-Bench, and more). Higher is better.
ClaudeLeaderClaude Fable 564.9
Cursor

CursorBench 3.1

Ambiguous, multi-file tasks from real Cursor sessions that test codebase understanding, bugfinding, planning, and code review.
ClaudeLeaderClaude Fable 572.9%

FrontierCode Diamond

The 50 hardest FrontierCode tasks: the toughest production-code problems, graded on whether maintainers would merge the patch. Scores stay low by design. Higher is better.
ClaudeLeaderClaude Opus 4.813.4%

FrontierCode Main

Cognition's test of whether a model writes code maintainers would actually merge, not just code that passes tests. Main is the 100 hardest of 150 tasks. Higher is better.
ClaudeLeaderClaude Fable 546.3%

Top AI coding agents

  • HermesSkills, Integrations & Self-hosting for Hermes
  • Claude CodeSkills, Plugins & MCP Servers for Claude Code
  • CodexSkills & MCP Servers for OpenAI Codex
  • OpenClawSkills & Automation for OpenClaw