TCSExam
Version Full
Agents prove theorems, write solvers, build mathematical constructions and address open conjectures in theoretical computer science.
The Full leaderboard reports seven models on 100 tasks. Its index combines four track scores and can exceed 100 points.
- Made by
- Moonshot AI
- Released
- November 2026
- Kind
- AI research tasks
- Level
- Research frontier
- Scored
- Checked automatically
- Answers
- Agent task
- Tasks
- All public
- Measures
- Capability
- Language
- English
- Interaction
- Agentic
- Environment
- Code sandbox
- Items from
- Written by experts
- Refresh
- One fixed set
- Run by
- A host
- In views
- Agents
- Strongest model tested
- GPT-6 Astra #4 on the capability index
- Last model added
- October 2026
- Scores refreshed
- November 21st, 2026
GPT-6 Astra leads with 71.9
Best score per model from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best.
- 1 GPT-6 Astra 71.9
- 2 Kimi K3.1 Pro 57.4
- 3 Kimi K3 46.8
- 4
DeepSeek V4.1 Flash 1017 46.0
- 5 GLM-5.3 45.5
- 6
Qwen3.8 Max (0902) 44.8
- 7 Grok 4.6 40.5
Source: Official TCSExam leaderboard
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | GPT-6 Astra | 71.9 | Sep 2026 | Official leaderboard | |
| 2 | Kimi K3.1 Pro | 57.4 | Oct 2026 | Official leaderboard | |
| 3 | Kimi K3 | 46.8 | Jul 2026 | Official leaderboard | |
| 4 | | 46.0 | Oct 2026 | Official leaderboard | |
| 5 | GLM-5.3 | 45.5 | Aug 2026 | Official leaderboard | |
| 6 | | 44.8 | Sep 2026 | Official leaderboard | |
| 7 | Grok 4.6 | 40.5 | Aug 2026 | Official leaderboard |
Rankings by part
We read 4 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Theorem, Solver and Construction. GPT-6 Astra leads 3 of them, and DeepSeek V4.1 Flash 1017 leads the other one. Arrows show how far a model moved from its overall rank among the models we track.
Theorem
7 models we track have a score here.
- 1 GPT-6 Astra 70.0
- 2 Kimi K3.1 Pro 38.3
- 3 Kimi K3 11.7
- 4
DeepSeek V4.1 Flash 1017 6.7
- 5 Grok 4.6 ↑2 5.0
- 6 GLM-5.3 ↓1 5.0
- 7
Qwen3.8 Max (0902) ↓1 5.0
Solver
7 models we track have a score here.
- 1 GPT-6 Astra 64.7
- 2 Kimi K3.1 Pro 54.1
- 3
DeepSeek V4.1 Flash 1017 ↑1 53.1
- 4
Qwen3.8 Max (0902) ↑2 49.2
- 5 GLM-5.3 49.1
- 6 Kimi K3 ↓3 48.7
- 7 Grok 4.6 43.8
Construction
7 models we track have a score here.
- 1 GPT-6 Astra 105.0
- 2 Kimi K3.1 Pro 98.6
- 3 GLM-5.3 ↑2 97.7
- 4 Kimi K3 ↓1 95.5
- 5
Qwen3.8 Max (0902) ↑1 95.1
- 6
DeepSeek V4.1 Flash 1017 ↓2 91.8
- 7 Grok 4.6 86.2
Open
7 models we track have a score here.
- 1
DeepSeek V4.1 Flash 1017 ↑3 5.6
- 2 Kimi K3.1 Pro 1.1
- 3 Kimi K3 0.0
- 4 Grok 4.6 ↑3 0.0
- 5 GLM-5.3 0.0
- 6
Qwen3.8 Max (0902) 0.0
- 7 GPT-6 Astra ↓6 0.0
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official TCSExam leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official TCSExam leaderboard
No lab model card reports it yet
We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.
Versions
More in AI research
- PaperBench In use Agent replicates ICML 2024 papers from scratch. LLM judge grades against author-written hierarchical rubrics.
Qwen 3.8 Max 93%
- MLS-Bench In use Agent must invent one algorithmic change in an ML research scaffold whose gain transfers across settings. Scored relative to human baselines. Claude Fable 5.1 50.3%
- PostTrainBench 1.2 In use CLI agent gets a base model, one H100 and a time budget to post-train it for a target benchmark. Scored on normalized gain. Claude Fable 5.1 44.6%
- MLE-bench In use Kaggle-style ML engineering competitions. Agent trains models and is scored against competition leaderboards. GPT-6.2 Astra 96.1%
- PMPP-Hard In use Agents write GPU kernels for 69 tasks that must be correct, and on 48 of them also run within 1.25 times the reference time. The score is the area under the solved-tasks curve over a 10-hour budget, out of 690 task-hours. Kimi K3 0.71
- KernelBench In use Replace PyTorch operators with faster custom GPU kernels. Scored on correctness and speedup.
MiniMax-M3 28.8%
Data audit
All 9 checks pass. Every benchmark
Catalog record 2 pass
- Pass Task access All tasks are public.
- Pass Confidence Checked against a primary source.
Scores 3 pass 2 notes
- Pass Independent runs Scores from the official leaderboard (7 entries).
- Pass Models scored 7 models we track.
- Pass Newest score The newest is from October 2026, for Kimi K3.1 Pro.
- Note Frontier models 1 of the top 10 has a score, which is fine for a retired test. The strongest tested is GPT-6 Astra (#4).
- Note Headroom Scored as an index, with no fixed ceiling.
Leaderboard import 3 pass
- Pass Importer We read the board on moonshotai.github.io: 7 rows for models we track at the last read.
- Pass Last read November 21st, 2026.
- Pass Read worked The last read worked.
Consistency 1 pass 1 note
- Note Lab claims No model has both a lab number and an independent run.
- Pass Known issues No published audit found problems.