SShogo Bench

DeepResearch Bench

PhD-level research reports with citations, graded for quality and citation accuracy.

Agent Index weight
5%
in the Knowledge-work agents pillar
Models tested
24
from the benchmark’s leaderboard and reports
Normalised for the index
0 → 0, 1 → 100
Official site

Results

#ModelScoreNormalisedHarnessCost/taskDate
1Claude Opus 4.6 Supported
Anthropic
55.3%source——
2Claude Sonnet 4.6 Supported
Anthropic
54.9%source——
3Claude Opus 4.5 (latest) Supported
Anthropic
54.8%source——
4GPT-5.5 Supported
OpenAI
54.0%source——
5Claude Sonnet 4.5 (latest) Supported
Anthropic
52.6%source——
6Claude Opus 4.8 Supported
Anthropic
50.2%source——
7Gemini 3 Flash Preview Supported
Google
49.8%source——
8GPT-5 Supported
OpenAI
49.6%source——
9Claude Opus 4.1 (latest) Supported
Anthropic
48.3%source——
10Gemini 3.1 Pro Preview Supported
Google
47.8%source——
11Grok 4 Supported
xAI
47.3%True——
12Claude-Opus-4 Supported
Anthropic
46.8%source——
13Claude Sonnet 4 Supported
Anthropic
46.6%source——
14Gemini 3 Pro Preview Supported
Google
46.3%source——
15Claude Haiku 4.5 Estimated
Anthropic
45.5%source——
16o3 Supported
OpenAI
45.2%source——
17Claude 3.7 Sonnet Supported
Anthropic
43.6%source——
18Gemini 2.5 Pro Supported
Google
42.8%source——
19GPT-5.1 Supported
OpenAI
42.8%source——
20GPT-5.2 Supported
OpenAI
41.1%source——
21Gemini 3.1 Flash Lite
Google
37.3%source——
22GPT-5.4 mini
OpenAI
36.3%source——
23DeepSeek-R1 Estimated
DeepSeek
35.1%source——
24GPT-5.4 Supported
OpenAI
35.1%source——

For models without a DeepResearch Bench result, the index estimates one from Epoch’s Capabilities Index using a logistic fit on 24 models (residual SD 0.24 on the logit scale). Estimates are labelled and widen the model’s range.