Skip to main content

Benchmark profile

Vals ProofBench (ProofBench)

Vals AI automated theorem-proving benchmark.

Data verified

How BenchLM shows ProofBench

BenchLM mirrors the public Vals AI ProofBench leaderboard captured from https://www.vals.ai/benchmarks/proof_bench and updated by Vals on July 17, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

ProofBench is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

51 Vals rows1 task viewsprivate datasetTasks: OverallDisplay only

ProofBench score on ProofBench — July 17, 2026

BenchLM mirrors the published proofbench score view for ProofBench. GPT-5.6 Sol leads the public snapshot at 77.00% , followed by Claude Fable 5 (77.00%) and GPT-5.6 Terra (71.00%). BenchLM does not use these results to rank models overall.

51 modelsExternal benchmark mirrorsCurrentDisplay onlyUpdated July 17, 2026

ProofBench score table (51 models)

Score
1
77.00%
2
77.00%
3
71.00%
4
AristotleHarmonic
71.00%
5
Kimi K3Moonshot AI
70.00%
6
69.00%
7
66.00%
8
GPT-5.4OpenAI
56.00%
9
54.00%
10
54.00%
11
GPT-5.5OpenAI
50.00%
12
50.00%
13
45.00%
14
39.00%
16
GLM 5.2Zhipu AI
35.00%
17
Grok 4.5SpaceXAI
30.00%
18
29.00%
19
26.00%
21
24.00%
22
GLM 5.1Zhipu AI
22.22%
23
21.00%
24
20.00%
25
MiniMax M3MiniMax
19.00%
27
GPT-5OpenAI
18.00%
28
17.00%
29
Kimi K2.6Moonshot AI
16.00%
30
Mimo V2.5Xiaomi
16.00%
32
GPT-5.2OpenAI
15.00%
33
14.00%
34
12.00%
35
Grok 4.3SpaceXAI
11.00%
36
10.00%
37
10.00%
38
9.00%
39
9.00%
40
8.00%
41
DeepSeek V3p2Fireworks AI
8.00%
42
GLM 4.7Zhipu AI
6.00%
43
5.00%
45
4.00%
46
4.00%
47
3.00%
48
InklingThinkingmachines
2.00%
50
Laguna Xs.2Poolside
0.00%
51
Laguna M.1Poolside
0.00%

The published ProofBench snapshot places GPT-5.6 Sol first at 77.00%. The third row is 6.00 points behind. The broader top-10 range is 23.00 points, so the table still separates the published systems.

51 models have been evaluated on ProofBench. The benchmark falls in the External benchmark mirrors category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. ProofBench is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About ProofBench

Year

2026

Tasks

Automated theorem proving

Format

Accuracy score

Difficulty

Formal proof reasoning

BenchLM mirrors Vals ProofBench as a display-only math and proof benchmark.

BenchLM freshness & provenance

Version

ProofBench 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

FAQ

What does ProofBench measure?

Vals AI automated theorem-proving benchmark.

Which model leads the published ProofBench snapshot?

GPT-5.6 Sol currently leads the published ProofBench snapshot with 77.00% proofbench score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on ProofBench?

51 AI models are included in BenchLM's mirrored ProofBench snapshot, based on the public leaderboard captured on July 17, 2026.

Last updated: July 17, 2026 · mirrored from the public benchmark leaderboard

Choose a model with this week’s evidence

Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.

One email each week. Unsubscribe anytime.