TCSExam-Lite - Solver Engineering: leaderboard
TCSExam (Moonshot AI) evaluates model-agent systems on verifiable theoretical computer science research: Lean proofs of results from recent TCS papers, reusable solvers judged on hidden instances, explicit constructions verified in Lean, and open conjectures. Unlike PutnamBench or miniF2F, which formalize competition problems, its theorems come from research papers and its solver and construction tracks are scored against published best-known results. Agent harness, version and reasoning setting per model follow the paper's Table 9 (arXiv 2611.07962v1). Question access: task statements and checkers are public on GitHub; solver judging uses hidden instances. Lite suite: 20 of the Full tasks, five per track; Offline Solver protocol with public self-tests during development and one final hidden-instance evaluation. Kimi K3-relative solver score: each instance outcome standardized against repeated fixed Kimi K3 reference runs and passed through a Gaussian CDF, bounded 0-100.
Metric: Kimi K3-relative solver score (%, bounded 0-100). Source: moonshotai.github.io. Saturation forecast: Rough model projection: around 2026. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-6 Astra (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Codex 0.154.0; effort=max) | 63.03 | |
| 2 | Claude Opus 5.5 (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Claude Code 2.1.283; 1M-token context option; effort=max) | 60.48 | |
| 3 | DeepSeek V4.1 Flash (1017) (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Claude Code 2.1.257; effort=max) | 54.33 | |
| 4 | Qwen 3.8 Max (0902) (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Claude Code 2.1.257; effort=max) | 50.95 | |
| 5 | Kimi K3.1 Pro (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Kimi Code 0.43.0; effort=max) | 49.98 | |
| 6 | Kimi K3 (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Kimi Code 0.43.0; effort=max) | 48.91 | |
| 7 | GLM-5.3 (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Claude Code 2.1.257; effort=max) | 46.05 | |
| 8 | Grok 4.6 (Benchmark Sweep Run: source=TCSExam; suite=lite; harness=Grok Build 1.0.41; thinking on; effort=xhigh) | 40.95 |
Interactive version: theaggregate.ai/benchmark?slug=tcsexam-lite-solver-engineering · How It Works · Data refreshed daily, snapshot 2026-11-12.