Skip to content

Benchmark

BrowseComp

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

Models measured
84
Scores recorded
203
Sources
4
IQ weight
2.43

Score distribution

66.4%
Mean
74.7%
Median
21.8%
Std dev
35.2%
10th percentile
90.1%
90th percentile
0.6%
Min
0.6%92.2%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.43 for chat models, measured over 84 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1GPT-5.6 SolAtria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · in-house search/visit/python harness · 500 steps · discard-all context strategy· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.2%
2GPT-5.6 SolUltra effort· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.2%
3Kimi K3.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim92.1%
4GPT 5.2Long Context · 128K · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92%
5LongCat 2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91.9%
6GPT-6 Astra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
91.5%
7GPT-6 AstraOpenAI launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91.5%
8Kimi K3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
91.2%
9Kimi K3Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · in-house search/visit/python harness · 500 steps · discard-all context strategy· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91.2%
10Kimi K3context compaction at 300K · 1M context · max effort · temp 1.0 · top_p 1.0· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91.2%
11Claude Opus 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
90.8%
12Claude Opus 5Atria Dawn Preview official card / report Table 1 · published comparison run · model snapshots and reasoning tiers not specified unless stated · in-house search/visit/python harness · 500 steps · discard-all context strategy· Reported by Shanghai AI Laboratory· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.8%
13Claude Opus 5OpenAI launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.8%
14Claude Opus 5Opus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.8%
15GPT-5.6 Sol· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
90.4%
16GPT-5.6 SolOpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.4%
17GPT-5.6 SolOpenAI launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.4%
18GPT-5.6 SolOpus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.4%
19Kimi K3no context management · full 1M context · max effort · temp 1.0 · top_p 1.0· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.4%
20Inkling 2 Small· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90.2%
21GPT-5.5 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
90.1%
22GPT-5.5 Proxhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.1%
23GPT 5.2Long Context · 256K · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.89.8%
24GPT-5.4 Pro· Steel.dev Agent Leaderboards data89.3%
25GPT-5.4 ProOpenAI GPT-5.4 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.89.3%
26Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 25, high· OpenRouter benchmark exports89%
27Claude Opus 4.8multi-agent· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.5%
28Claude Fable 5· Steel.dev Agent Leaderboards data88%
29GPT-5.6 Terra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim87.5%
30GPT-5.6 TerraOpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.5%
31Claude Fable 5OpenAI launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.4%
32Claude Fable 5Opus 5 System Card capability summary · adaptive thinking · max effort unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.4%
33ERNIE 5.5 Flash Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim87.4%
34Claude Sonnet 5multi-agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.6%
35Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 25, high· OpenRouter benchmark exports86.3%
36Kimi K2.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
86.3%
37Kimi K2.6Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · Agent Swarm· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.3%
38Doubao Seed 2.1 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.2%
39Gemini 3.1 Pro Preview· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
85.9%
40Gemini 3.1 Pro PreviewFable/Mythos 5 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
41Gemini 3.1 Pro PreviewOpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
42Gemini 3.1 Pro PreviewOpus 4.8 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
43Gemini 3.1 Pro PreviewSearch + Python + Browse· Reported by Google DeepMind· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
44Doubao Seed 2.1 Turbo· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.9%
45Claude Sonnet 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
84.7%
46Claude Sonnet 5single agent · search/fetch/code · 10M token limit · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.7%
47Mistral Large 4· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim84.5%
48GPT 5.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
84.4%
49GPT 5.5Fable/Mythos 5 System Card capability summary · adaptive thinking · max effort · 5 trials unless noted· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.4%
50GPT 5.5OpenAI GPT-5.6 launch table· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.4%

Showing 1–50 of 203 results.

Data credits: Steel.dev Agent Leaderboards, MIT (licence) · 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. · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence)