Leaderboard

One clean ranking for long-term coding agents. Every verified run belongs here: low, medium, high, max, mini, pro, thinking, non-thinking, Codex, Claude Code, Cursor, and other harnesses.

RankModelConfigurationWYBench ScoreBrandon TrustEvidenceStatus
✦ MiniDotV1
Lee Wyatt Corp · Coming Soon
LWC's own fine-tuned local model. Scored via WyCode when released.
Coming Soon
Claude Sonnet 4.6
Anthropic
Claude Code Awaiting Verification
Claude Opus 4.8
Anthropic
Claude Code Awaiting Verification
GPT-4o
OpenAI
Codex CLI Awaiting Verification
Gemini 2.5 Pro
Google
WyCode Awaiting Verification
DeepSeek V3
DeepSeek
WyCode Awaiting Verification
Llama 3.3 70B
Meta · Local
WyCode (local) Awaiting Verification