Skip to content

Benchmark

BrowseComp-zh

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

Models measured
17
Scores recorded
21
Sources
2
IQ weight
2.12

Score distribution

61.8%
Mean
65%
Median
13.1%
Std dev
47.9%
10th percentile
72.7%
90th percentile
25%
Min
25%81.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.12 for chat models, measured over 17 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim81.3%
2Step 3.5 FlashStep 3.5 Flash official headline table · vanilla · with Context Manager· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.73.7%
3GLM 5GLM-5 official release table· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.72.7%
4Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim70.3%
5Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.70.3%
6Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim69.9%
7Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim69.5%
8LongCat-Flash-Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim69%
9Step 3.5 FlashStep 3.5 Flash official headline table · vanilla· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.66.9%
10GLM 4.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim66.6%
11DeepSeek Reasoner· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim65%
12DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim65%
13DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · commercial search API· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.65%
14Kimi K2 Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim62.3%
15Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim62.1%
16Qwen3 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim60.9%
17DeepSeek V3.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim49.2%
18MiniMax M2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it.48.5%
19DeepSeek V3.2 Exp· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim47.9%
20DeepSeek R1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim35.7%
21Step 3.5 FlashStep 3.5 technical report · Table 11 · 256K · no-tools baseline· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.25%

Data credits: Data by LLM Stats · 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.