List your company

Trajectory step classification on Trajectory nine-category label set

95Label AccuracyClaude Haiku 4.5
52.04863.19974.3585.501May 19, 2026

Evaluation results

MethodLinks
Claude Haiku 4.5Tier=frontier, Mode=aware2026.059594.79.287.92096.287.985.4
Gemma-4-26B-A4B-itTier=open, Mode=aware2026.0594.393.78.993.419.89576.574.3
Gemini 3.1 Flash LiteTier=frontier, Mode=aware2026.0592.792.59.291.119.793.785.682.4
Ministral-3-14BTier=open, Mode=aware2026.0590.590.28.976.91991.586.880.2
Qwen3.6-35B-A3BTier=open, Mode=aware2026.0589.589.714.692.623.3898780.5
Claude Sonnet 4.6Tier=frontier, Mode=aware2026.0588.588.69.394.219.690.386.579.6
gpt-5.4-miniTier=frontier, Mode=aware2026.0581.882.714.689.122.383.887.376.2
gpt-oss-20BTier=open, Mode=aware2026.0568.666.710.585.818.578.981.566
gpt-5.4-miniTier=frontier, Mode=blind2026.0561.561.510.68218.5———
Claude Sonnet 4.6Tier=frontier, Mode=blind2026.0561.361.39.590.918.4———
Claude Haiku 4.5Tier=frontier, Mode=blind2026.0556.958.68.485.916.9———
Gemini 3.1 Flash LiteTier=frontier, Mode=blind2026.0556.458.59.291.617.6———
Qwen3.6-35B-A3BTier=open, Mode=blind2026.0555.556.99.182.917———
Gemma-4-26B-A4B-itTier=open, Mode=blind2026.0554.2568.688.316.8———
gpt-oss-20BTier=open, Mode=blind2026.0553.8527.887.816.1———
Ministral-3-14BTier=open, Mode=blind2026.0553.756.18.977.716.5———