BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

ALE-Bench leaderboard

As of 6 Nov 2026, the top score on ALE-Bench is 2951.3 by GPT-6 Astra (OpenAI), ahead of GPT-6.1 Sol at 2743.8, across 131 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 3 Sep 2026). The model iterates on a heuristic-contest solution over hours, scored like a contestant. ALE-Bench is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • GPT-6 Astra holds the top ALE-Bench score at 2951.3 (Epoch AI Benchmarking Hub, 3 Sep 2026).
  • GPT-6.1 Sol is second on ALE-Bench at 2743.8 (Epoch AI Benchmarking Hub, 29 Sep 2026).
  • ALE-Bench: Long-horizon algorithm engineering on AtCoder heuristic contests (Epoch AI Benchmarking Hub).
  • 131 model configurations have a published ALE-Bench result on BenchLeader (BenchLeader, 6 Nov 2026).
  • ALE-Bench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Long-horizon algorithm engineering on AtCoder heuristic contests.

Published by
Epoch AI Benchmarking Hub
Category
Coding
Index weight
Reference only
Models
131
Data as of
6 Nov 2026
Newest result
26 Oct 2026

CC BY 4.0 — Epoch AI, ‘AI Benchmarking Hub’, epoch.ai/benchmarks. Mirrored boards credit their original publishers.

What the test looks like

The model iterates on a heuristic-contest solution over hours, scored like a contestant.

How it is scored

Contest performance rating, as published by ALE-Bench.

What to keep in mind

Rating scale like competitive programming; not a percent.

Top 15
  1. 1GPT-6 Astra (max)2951.3
  2. 2GPT-6.1 Sol (max)2743.8
  3. 3GPT-6.1 Terra (max)2608.8
  4. 4GPT-6 Sol (max)2462
  5. 5GPT-5.6 Sol (max)2176.9
  6. 6Claude Opus 5 (high)2164.6
  7. 7Claude Opus 5.5 (high)2147.4
  8. 8Claude Fable 5.1 (high)2143.2
  9. 9Gemini 4 Argon (high)2100.1
  10. 10Claude Fable 5 (high)2041.3
  11. 11GPT-5.6 Terra (max)1951.4
  12. 12GPT-5.5 (xhigh)1943.0
  13. 13Kimi K3.1 Pro (max)1932.8
  14. 14Claude Sonnet 5.5 (high)1819.1
  15. 15Gemini 4 Argon (low)1705.4
131 of 131
#
1GPT-6 Astramax2951.367.82026-09-03
2GPT-6.1 Solmax2743.867.32026-09-29
3GPT-6.1 Terramax2608.865.32026-10-26
4GPT-6 Solmax246263.32026-09-22
5GPT-5.6 Solmax2176.965.42026-07-09
6Claude Opus 5high2164.664.32026-07-24
7Claude Opus 5.5high2147.465.82026-09-22
8Claude Fable 5.1high2143.264.02026-09-01
9Gemini 4 Argonhigh2100.166.02026-09-30
10Claude Fable 5high2041.364.32026-06-09
11GPT-5.6 Terramax1951.460.72026-07-09
12GPT-5.5xhigh1943.064.32026-04-23
13Kimi K3.1 Promax1932.8–2026-10-23
14Claude Sonnet 5.5high1819.158.62026-09-28
15Gemini 4 Argonlow1705.4–2026-09-30
16Muse Spark 1.3xhigh1688.358.52026-09-02
17GPT-5.6 Lunamax1667.456.32026-07-09
18GPT-5.3-Codexxhigh1655.262.92026-02-05
19DeepSeek V4.1 Flash 1017max162259.22026-10-17
20GPT-5.4high160758.82026-03-05
21Muse Spark Open 816B A42Bxhigh1590.158.52026-10-15
22GPT-5.5medium1589.461.02026-04-23
23Claude Haiku 5.5high158755.82026-10-07
24GPT-6 Lunaxhigh1576.952.42026-09-22
25Claude Opus 4.8high1563.860.62026-05-28
26Kimi K3max1524.562.12026-07-16
27GPT-5.4medium1520.754.72026-03-05
28Grok 4.6xhigh1508.157.92026-08-12
29Claude Sonnet 5high1463.159.32026-06-30
30Claude Opus 4.8no reasoning1411.8–2026-05-28
31DeepSeek V4 Promax1403.257.02026-08-13
32Gemini 3 Flash1367.256.92025-12-17
33Claude Sonnet 4.6medium1327.3–2026-02-17
34Claude Opus 4.71323.063.12026-04-16
35GLM 5.3high1317.4–2026-08-14
36Grok 4.5high1309.356.92026-07-08
37DeepSeek V4 Flashmax1306.157.02026-07-31
38GPT-5.2-Codex1299.957.32025-12-18
39GPT-5.2high1293.556.02025-12-11
40Gemini 3.8 Flashhigh1269.762.02026-09-02
41GPT-5.2medium1249.856.22025-12-11
42GPT-5.1-Codex1244.9–2025-11-12
43MiMo-V2.6-Flash1244.255.02026-09-21
44Mistral Large 41213.252.92026-10-06
45GPT-5.1-Codex-Maxmax1208.8–2025-11-19
46GPT-5.1high1192.257.72025-11-13
47Qwen3.7max1189.460.32026-05-19
48GPT-5.4 Minihigh1188.652.12026-03-17
49Gemini 3 Pro1176.860.22025-11-18
50GPT-5high1162.557.92025-08-07
51Gemini 3.1 Pro1160.660.52026-02-19
52MiMo-V2.6-Pro1158.157.32026-09-21
53Grok 4.201150.3–2026-02-17
54GPT-5.5no reasoning1127.652.32026-04-23
55Kimi K2.61092.755.22026-04-20
56DeepSeek V4.1 Flashmax1092.359.02026-09-09
57GPT-5.4no reasoning1086.053.02026-03-05
58GLM 5.2high1047.0–2026-06-16
59Claude Opus 4.51025.457.92025-11-24
60GLM 5.2max1010.259.52026-06-16

More coding benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on ALE-Bench

Cite as: BenchLeader, “ALE-Bench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/ale_bench.

ALE-Bench: questions

What does ALE-Bench measure?
The model iterates on a heuristic-contest solution over hours, scored like a contestant. Scores are reported in score; higher is better.
What is the top score on ALE-Bench?
As of 6 Nov 2026, the highest score on ALE-Bench is 2951.3 by GPT-6 Astra (result dated 3 Sep 2026), ahead of GPT-6.1 Sol at 2743.8. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
Which AI model leads ALE-Bench?
GPT-6 Astra leads ALE-Bench with 2951.3 as of 6 Nov 2026 (result dated 3 Sep 2026), ahead of GPT-6.1 Sol at 2743.8.
How many models have ALE-Bench results?
131 model configurations have a ALE-Bench result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
Who runs ALE-Bench and how often is it updated?
ALE-Bench is published by Epoch AI Benchmarking Hub. BenchLeader re-reads the published results once a day and records the date each result was published.
Does ALE-Bench count toward the BenchLeader Index?
No. ALE-Bench is shown for reference but left out of the composite index.