OpenCodePapers

agents-on-posttrainbench-v1-1

Agents
Dataset Link
Results over time
Click legend items to toggle metrics. Hover points for model names.
Leaderboard
PaperCodeAccuracyModelNameReleaseDate
Gemini 4 Argon: our next era of frontier intelligence✓ Link45.3Gemini 4 Argon (OpenCode)2026-09-30
PostTrainBench Leaderboard✓ Link41.79Claude Fable 5 (1M, Max) (Claude Code)2026-06-09
GLM-5.3: Frontier Coding with Emergent Cyber Capabilities✓ Link39.8GLM-5.3 (Claude Code)2026-08-14
PostTrainBench Leaderboard✓ Link36.23GPT-5.6 Sol (Max) (Codex CLI)2026-07-09
Hy4 preview✓ Link35.6Hy4 preview2026-08-28
PostTrainBench Leaderboard✓ Link35.04Claude Opus 5 (Claude Code)2026-07-24
PostTrainBench Leaderboard33.84Claude Opus 4.8 (High) (Claude Code)2026-05-28
PostTrainBench Leaderboard✓ Link31.96Kimi K3 (Claude Code)2026-07-27
PostTrainBench Leaderboard✓ Link31.70GLM-5.2 (Max) (Claude Code)2026-06-16
PostTrainBench Leaderboard28.56Claude Opus 4.7 (xHigh) (Claude Code)2026-04-16
PostTrainBench Leaderboard27.23GPT-5.5 (xHigh) (Codex CLI)2026-04-23
PostTrainBench Leaderboard23.45Grok 4.5 (High) (Cursor CLI)2026-07-08
PostTrainBench Leaderboard21.99Gemini 3.1 Pro (OpenCode)2026-02-19
PostTrainBench Leaderboard19GPT-5.4 (High) (Codex CLI)2026-03-05