Skip to content

Benchmark

DeepResearch Bench II

Agentic benchmarks measure how well models carry out tasks that take several steps. Scores are percentages. Higher is better.

Models measured
6
Scores recorded
6
Sources
1
IQ weight
0.44

Score distribution

50.8%
Mean
51.3%
Median
2.4%
Std dev
49.2%
10th percentile
54.1%
90th percentile
46.6%
Min
46.6%54.1%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.44 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Claude Opus 5Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.54.1%
2GLM 5.3Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.52.7%
3Kimi K3Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.3%
4GPT-5.6 SolAtria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.50.7%
5Qwen3.8 MaxAtria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.49.2%
6DeepSeek V4 Pro 0813Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · Claude Code · identical search tools · four-hour task limit · GPT-5.5 judge · three independent scoring passes· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.46.6%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links.