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
101GLM 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
75.9%
102GLM 5GLM-5 official release table · with context management· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.75.9%
103Agents A1· Steel.dev Agent Leaderboards data75.5%
104MiniMax M2.5search/tool figure · with context· Reported by MiniMax· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.75.1%
105Kimi K2.5single agent · context management· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.74.9%
106Claude Sonnet 4.6launch-page figure · web search/fetch + code · no thinking· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.74.7%
107Claude Sonnet 4.6single agent · 10M total sampled tokens · no thinking · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.74%
108GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, perplexity, server-tool, 25, xhigh· OpenRouter benchmark exports74%
109DeepSeek V4 FlashDeepSeek V4 official model card · Think Max· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.73.2%
110LongCat-Flash-Thinking· Steel.dev Agent Leaderboards data
Values from 2 sources
73.1%
111Ling 3.0 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim72.2%
112Claude Sonnet 4.6single agent · 3M total sampled tokens · no thinking · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.69.7%
113DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-0731, parallel, server-tool, 25, high· OpenRouter benchmark exports69.1%
114Qwen3.5 397B A17BQwen3.5 release/model card · simple context folding · 256K· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.69%
115Step 3.5 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
69%
116Step 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.69%
117Step 3.5 FlashStep 3.5 technical report · Table 17 · 200-item subset · multi-agent · 721 real steps· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.68.5%
118GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, exa, server-tool, 25, xhigh· OpenRouter benchmark exports68.4%
119GLM 5.1GLM-5.1 official release table · without context management· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.68%
120DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · Discard-all context management· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.67.6%
121Doubao Seed 1.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim67.6%
122GLM 4.7· Steel.dev Agent Leaderboards data
Values from 2 sources
67.5%
123DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-0731, exa, server-tool, 25, high· OpenRouter benchmark exports67.4%
124Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 5, high· OpenRouter benchmark exports66.5%
125Step 3.5 FlashStep 3.5 technical report · Table 17 · 200-item subset · discard-all · 302 real steps· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.66%
126GPT 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
65.8%
127GPT 5.2OpenAI 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.65.8%
128GPT 5.2standard tool-use setup · xhigh· Reported by OpenAI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.65.8%
129GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 5, high· OpenRouter benchmark exports65.2%
130Claude Sonnet 4.6single agent · 1M total sampled tokens · no thinking · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.64.7%
131Qwen3.5 122B-A10B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
63.8%
132GLM 5GLM-5 official release table · without context management· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.62%
133MiniMax M2.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
62%
134Qwen3.5 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
61%
135Qwen3.5 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
61%
136Kimi K2.5single agent · no context management· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.60.6%
137GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 5, high· OpenRouter benchmark exports60.2%
138DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · Summary context management · average steps 140→364· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.60.2%
139Kimi K2 Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
60.2%
140GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 5, high· OpenRouter benchmark exports60.1%
141Gemini 3 Pro Preview· Steel.dev Agent Leaderboards data59.2%
142GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 5, high· OpenRouter benchmark exports59%
143MiMo V2 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
58.3%
144MiMo V2 FlashMiMo-V2-Flash final evaluation · with context management· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.58.3%
145Step 3.5 FlashStep 3.5 technical report · Table 17 · 200-item subset · keep-first-and-last-K · 244 real steps· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.58%
146Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 5, high· OpenRouter benchmark exports57.8%
147GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, perplexity, server-tool, 5, xhigh· OpenRouter benchmark exports57%
148Step 3.5 FlashStep 3.5 technical report · Table 17 · 200-item subset · summary · 131 real steps· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.57%
149GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, parallel, server-tool, 25, xhigh· OpenRouter benchmark exports56%
150GPT 5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim
Values from 2 sources
54.9%

Showing 101–150 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)