BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

DTBench leaderboard

As of 6 Nov 2026, the top score on DTBench is 98.9% by Claude Opus 5.5 (Anthropic), ahead of Claude Fable 5 at 98.4%, across 230 model configurations with a published result (Epoch AI Benchmarking Hub, result dated 22 Sep 2026). Decision problems where the model's answer preference is compared with decision-theoretic expert answers. DTBench is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5.5 holds the top DTBench score at 98.9% (Epoch AI Benchmarking Hub, 22 Sep 2026).
  • Claude Fable 5 is second on DTBench at 98.4% (Epoch AI Benchmarking Hub, 9 Jun 2026).
  • DTBench: Decision-theory dilemmas scored against expert answers (Epoch AI Benchmarking Hub).
  • 230 model configurations have a published DTBench result on BenchLeader (BenchLeader, 6 Nov 2026).
  • DTBench does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Decision-theory dilemmas scored against expert answers.

Published by
Epoch AI Benchmarking Hub
Category
Reasoning
Index weight
Reference only
Models
230
Data as of
6 Nov 2026

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

What the test looks like

Decision problems where the model's answer preference is compared with decision-theoretic expert answers.

How it is scored

Accuracy, as published by DTBench.

What to keep in mind

Philosophy-of-decision framing; a large model set.

Top 15
  1. 1Claude Opus 5.5 (max)98.9%
  2. 2Claude Fable 5 (max)98.4%
  3. 3Claude Opus 5 (high)97.9%
  4. 4Claude Opus 5 (xhigh)97.9%
  5. 5Claude Fable 5.197.6%
  6. 6Claude Opus 5 (max)97.6%
  7. 7Claude Fable 5.1 (xhigh)97.6%
  8. 8Gemini 4 Argon (low)97.6%
  9. 9GPT-6 Astra (xhigh)97.3%
  10. 10Claude Opus 5.5 (xhigh)97.3%
  11. 11Gemini 4 Argon (high)97.3%
  12. 12Claude Sonnet 5.5 (max)97.3%
  13. 13GPT-6 Sol (max)97.3%
  14. 14Grok 4.6 (xhigh)97.3%
  15. 15Gemini 4 Argon (medium)97.3%
230 of 230
#
1Claude Opus 5.5max98.9%69.32026-09-22
2Claude Fable 5max98.4%65.32026-06-09
3Claude Opus 5high97.9%64.32026-07-24
4Claude Opus 5xhigh97.9%60.22026-07-24
5Claude Fable 5.197.6%66.42026-09-01
6Claude Opus 5max97.6%65.42026-07-24
7Claude Fable 5.1xhigh97.6%65.42026-09-01
8Gemini 4 Argonlow97.6%–2026-09-30
9GPT-6 Astraxhigh97.3%66.82026-09-03
10Claude Opus 5.5xhigh97.3%66.32026-09-22
11Gemini 4 Argonhigh97.3%66.02026-09-30
12Claude Sonnet 5.5max97.3%65.42026-09-28
13GPT-6 Solmax97.3%63.32026-09-22
14Grok 4.6xhigh97.3%57.92026-08-12
15Gemini 4 Argonmedium97.3%–2026-09-30
16Claude Fable 5xhigh97.3%–2026-06-09
17GPT-6 Astra97.1%68.22026-09-03
18GPT-6 Astramedium97.1%65.12026-09-03
19Claude Fable 5.1medium97.1%62.32026-09-01
20Gemini 3.1 Promedium97.1%–2026-02-19
21Gemini 3.7 Flashhigh96.8%62.22026-08-13
22GPT-6.1 Solmax96.5%67.32026-09-29
23GPT-6 Astrahigh96.5%67.02026-09-03
24Claude Opus 5.5high96.5%65.82026-09-22
25Claude Fable 5high96.5%64.32026-06-09
26Claude Fable 5.1high96.5%64.02026-09-01
27Claude Opus 5.5medium96.5%62.02026-09-22
28Muse Spark 1.396.5%57.72026-09-02
29Grok 4.5high96.5%56.92026-07-08
30GPT-5.5xhigh96.0%64.32026-04-23
31Claude Sonnet 5.5xhigh96.0%63.42026-09-28
32Grok 4.7xhigh96.0%58.32026-09-21
33GPT-5.5 Proxhigh96.0%–2026-04-23
34GPT-5.6 Solpro mode96.0%–2026-07-09
35GPT-5.6 Solxhigh95.7%64.22026-07-09
36Claude Fable 5medium95.7%–2026-06-09
37Gemini 3.8 Flash95.7%54.42026-09-02
38GPT-5.6 Solmax95.5%65.42026-07-09
39GPT-5.6 Solhigh95.5%63.62026-07-09
40Gemini 3.6 Flashhigh95.5%61.22026-07-21
41Gemini 3.1 Prohigh95.5%60.12026-02-19
42Claude Opus 5.5low95.5%59.52026-09-22
43Muse Wonder95.2%63.92026-10-12
44GPT-5.6 Solmedium95.2%61.52026-07-09
45Claude Haiku 5.5max95.0%57.02026-10-07
46GPT-6 Astralow94.9%62.72026-09-03
47Gemini 3.8 Flashmedium94.9%60.72026-09-02
48Claude Opus 4.8max94.9%60.22026-05-28
49Claude Opus 5medium94.9%57.42026-07-24
50Gemini 3.8 Flashlow94.9%56.22026-09-02
51Gemini 3.1 Prolow94.9%–2026-02-19
52Gemini 3.5 Flashhigh94.7%62.52026-05-19
53Muse Spark 1.2xhigh94.7%61.72026-08-05
54Gemini 3.7 Flashmedium94.7%60.32026-08-13
55Claude Opus 4.7max94.7%59.52026-04-16
56GPT-5.4xhigh94.4%63.22026-03-05
57Muse Spark 1.1high94.4%–2026-07-09
58Claude Fable 5.1low94.1%59.32026-09-01
59Claude Sonnet 5.5high94.1%58.62026-09-28
60GPT-5.6 Sollow93.9%60.42026-07-09

More reasoning benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on DTBench

Cite as: BenchLeader, “DTBench leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/dtbench.

DTBench: questions

What does DTBench measure?
Decision problems where the model's answer preference is compared with decision-theoretic expert answers. Scores are reported in percent of tasks solved; higher is better.
What is the top score on DTBench?
As of 6 Nov 2026, the highest score on DTBench is 98.9% by Claude Opus 5.5 (result dated 22 Sep 2026), ahead of Claude Fable 5 at 98.4%. Scores come from Epoch AI Benchmarking Hub and are re-read daily.
Which AI model leads DTBench?
Claude Opus 5.5 leads DTBench with 98.9% as of 6 Nov 2026 (result dated 22 Sep 2026), ahead of Claude Fable 5 at 98.4%.
How many models have DTBench results?
230 model configurations have a DTBench result on BenchLeader, all taken from Epoch AI Benchmarking Hub.
Who runs DTBench and how often is it updated?
DTBench 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 DTBench count toward the BenchLeader Index?
No. DTBench is shown for reference but left out of the composite index.