Today for AI
HOT RADAR
llmHEAT 15.7°

AI CURATED HOT STORY · 11/9/2026

Tsinghua and Moonshot AI Launch TCSExam: GPT-6 Astra Proves 70% of Theorems but Solves No Open Conjectures

1 reports archived1 independent sourcesupdated 11/9/2026, 02:45:52
Synthesis & Latest Updates
1 Sources Cross-Validated

Tsinghua University and Moonshot AI released TCSExam, a 100-task expert-reviewed benchmark spanning Lean theorem proving, solver engineering, record-setting constructions, and 30 open conjectures, with every submission verified by the Lean kernel or a task checker. GPT-6 Astra fully proved 70.0% of theorems without supplied lemmas, versus 38.3% for Kimi K3.1 Pro and at most 11.7% for six other agents, yet the same model solved none of the 30 open TCS conjectures; the few accepted conjecture submissions merely adapted counterexamples and proofs OpenAI had already published.

LATEST/Tsinghua University and Moonshot AI released TCSExam, a 100-task expert-reviewed benchmark spanning Lean theorem proving, solver engineering, record-setting constructions, and 30 open conjectures, with every submission verified by the Lean kernel or a task checker. GPT-6 Astra fully proved 70.0% of theorems without supplied lemmas, versus 38.3% for Kimi K3.1 Pro and at most 11.7% for six other agents, yet the same model solved none of the 30 open TCS conjectures; the few accepted conjecture submissions merely adapted counterexamples and proofs OpenAI had already published.

Hourly heat curve

2 fully observed hours
Fewer than 3 fully observed hours; trend not plotted yet.

Coverage timeline

Total 1 reports · Latest first
  1. Rohan Paul (@rohanpaul_ai)T2·78 pts

    Tsinghua and Moonshot AI Launch TCSExam: GPT-6 Astra Proves 70% of Theorems but Solves No Open Conjectures

    • TCSExam has 100 expert-reviewed tasks across four tracks—Lean proving, solver engineering, record-setting constructions, and 30 open conjectures—all verified by the Lean kernel or a checker
    • Without supplied lemmas, GPT-6 Astra fully proved 70.0% of theorems, Kimi K3.1 Pro 38.3%, and six other agents at most 11.7%
    • The same model solved none of the 30 open TCS conjectures; the few accepted submissions only adapted counterexamples and proofs OpenAI had already published