organizational research article

PostTrainBench v1.2: More reliable scores and verdicts

PostTrainBench team · Published 2026-09-30

Evidence from this source

Results and research linked to this document
Type	Evidence
Result	PostTrainBench · Claude Fable 5.1 + Opus 5 GPQA fallback / Claude Code, Max Source details New agents and recomputed results
Result	PostTrainBench · Claude Opus 5.5 / Claude Code, Max Source details New agents and recomputed results
Result	PostTrainBench · GPT-6 Astra / Codex CLI, Max Source details New agents and recomputed results
Result	PostTrainBench · Locus (Intology, powered by Opus 5) Source details New agents and recomputed results
Result	PostTrainBench · Fable 5 + Opus 4.8 Max GPQA fallback Source details New agents and recomputed results
Result	PostTrainBench · GLM 5.3 Flash / Claude Code, Max Source details New agents and recomputed results
Result	PostTrainBench · Claude Opus 5 / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · GLM 5.3 / Claude Code, Max Source details New agents and recomputed results
Result	PostTrainBench · Kimi K3, 1M context / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · GPT-5.6 Sol, Max reasoning / Codex CLI Source details New agents and recomputed results
Result	PostTrainBench · Claude Opus 4.8, Max reasoning / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · Claude Opus 4.8, High reasoning / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · GLM 5.2, Max reasoning / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · Claude Opus 4.7, xHigh reasoning / Claude Code Source details New agents and recomputed results
Result	PostTrainBench · GPT-5.5, xHigh reasoning / Codex CLI Source details New agents and recomputed results
Result	PostTrainBench · Grok 4.5, High reasoning / Cursor CLI Source details New agents and recomputed results
Result	PostTrainBench · Gemini 3.1 Pro / OpenCode Source details New agents and recomputed results
Result	PostTrainBench · GPT-5.4, High reasoning / Codex CLI Source details New agents and recomputed results
Benchmark	PostTrainBench Source details Version changes, new agents and Fable 5.1 fallback

Document tracking details

Source ID: src-ptb-v12-release
Source type: Primary
Original URL: https://posttrainbench.com/blog/posttrainbench-1-2/
Retrieved URL: https://posttrainbench.com/blog/posttrainbench-1-2/
Last updated by publisher: Not reported
Access status: available
Retrieved: 2026-10-02T02:03:46.351552Z
Document fingerprint: 4af5f190d734779181499888ad0ee7ba2b5ce0b753b19ef0a9db0802fff580b5
Based on original bytes. Used to detect changes to the source.

Availability checks

2026-10-02T02:03:46.351552Z · success: Retained original source bytes; remote code inspected as text only.