Theorem Proving on ProverBench Number Theory
25Solved ProblemsDreamProver
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DreamProverMethod Category=Lemma Learning System, Backbone=Gemini 2.5 Pro2026.04 | 25 | 62.5 | |
| DreamProverMethod Category=Lemma Learning System, Backbone=Gemini 3.1 Pro2026.04 | 25 | 62.5 | |
| DreamProverMethod Category=Lemma Learning System, Backbone=GPT-5.3-Codex2026.04 | 21 | 52.5 | |
| HilbertMethod Category=Agentic System, Backbone=Gemini 3.1 Pro2026.04 | 16 | 40 | |
| Gemini 3.1 ProMethod Category=Proprietary LLMs2026.04 | 13 | 32.5 | |
| HilbertMethod Category=Agentic System, Backbone=GPT-5.3-Codex2026.04 | 13 | 32.5 | |
| GPT-5.3-CodexMethod Category=Proprietary LLMs2026.04 | 12 | 30 | |
| Goedel-Prover-V2-32BMethod Category=Open-source LLMs2026.04 | 12 | 30 | |
| Claude 4.6 OpusMethod Category=Proprietary LLMs2026.04 | 11 | 27.5 | |
| HilbertMethod Category=Agentic System, Backbone=Gemini 2.5 Pro2026.04 | 11 | 27.5 | |
| DeepSeek-Prover-V2-7BMethod Category=Open-source LLMs2026.04 | 10 | 25 | |
| Goedel-Prover-V2-8BMethod Category=Open-source LLMs2026.04 | 10 | 25 | |
| Gemini 2.5 ProMethod Category=Proprietary LLMs2026.04 | 5 | 12.5 |