Skip to content

Benchmark

DeepSearchQA

Scores are percentages. Higher is better.

Models and systems measured
24
Scores recorded
68
Sources
3
IQ weight
2.70

Score distribution

69.3%
Mean
73.6%
Median
18.7%
Std dev
42%
10th percentile
91%
90th percentile
22%
Min
22%95%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.70 for chat models, measured over 23 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
1Kimi K3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim95%
2Kimi K3max 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.95%
3Claude Opus 4.8· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim93.1%
4GPT-5.6 Solmax · 900 questions · common browser search/open/find harness · semantic answer matching· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93.1%
5Kimi K2.6Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context · no context management; over-length tasks counted failed· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.5%
6Claude Opus 4.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91.3%
7Claude Sonnet 4.6multi-agent orchestrator · 3M tokens per agent · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91.1%
8Hy3 (Deep Infra)· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91%
9Hy3 (Deep Infra)Hy3 official Benchmark Appendix · highest reasoning tier unless noted· Reported by Tencent Hunyuan· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91%
10Claude Opus 5max · 900 questions · common browser search/open/find harness · semantic answer matching· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.4%
11Muse Spark 1.3max · 900 questions · common browser search/open/find harness · semantic answer matching· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90.3%
12Muse Spark 1.3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim89.4%
13Claude Sonnet 4.6single agent · 10M total tokens · search/fetch/code · adaptive thinking · max effort· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.89.2%
14Claude Opus 4.7search/fetch + code · adaptive thinking · max effort · 10M total tokens · compaction at 200K· Reported by Anthropic· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.89.1%
15GPT 5.5xhigh · browser search/open/find · gpt-oss-120b answer extraction· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.87.8%
16MiMo V2 Pro· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86.7%
17MiMo V2 ProXiaomi MiMo-V2-Pro release table · search-agent setting · harness unspecified· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.7%
18Muse Spark 1.2xhigh · 900 questions · common browser search/open/find harness · semantic answer matching· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.9%
19Inkling 2 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.3%
20Muse Spark 1.1Meta Model API · xhigh · browser search/open/find · gpt-oss-120b answer extraction· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.9%
21Claude Opus 4.8max · browser search/open/find · gpt-oss-120b answer extraction· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.3%
22Kimi K2.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83%
23Beam 1.0 501B A23B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim80.5%
24Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim77.1%
25Kimi K2.5setting not named by the source· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.77.1%
26Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 25, high· OpenRouter benchmark exports77%
27Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 25, high· OpenRouter benchmark exports76.5%
28GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 25, high· OpenRouter benchmark exports75%
29Muse Spark· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim74.8%
30Muse Glimmer 30B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim74.6%
31Muse Glimmer 30Bhigh reasoning· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.74.6%
32GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 25, high· OpenRouter benchmark exports73.6%
33GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 25, high· OpenRouter benchmark exports73.6%
34GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, perplexity, server-tool, 25, xhigh· OpenRouter benchmark exports73%
35Mercury 3· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim71.3%
36Gemini 3.1 Pro Previewhigh · browser search/open/find · gpt-oss-120b answer extraction· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.71.3%
37Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, exa, server-tool, 25, high· OpenRouter benchmark exports70.5%
38DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-0731, perplexity, server-tool, 25, high· OpenRouter benchmark exports69%
39GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 25, high· OpenRouter benchmark exports69%
40GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, exa, server-tool, 25, xhigh· OpenRouter benchmark exports67%
41GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, native, server-tool, 5, high· OpenRouter benchmark exports67%
42Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, perplexity, server-tool, 5, high· OpenRouter benchmark exports64.7%
43DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-0731, parallel, server-tool, 25, high· OpenRouter benchmark exports64%
44GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, parallel, server-tool, 25, xhigh· OpenRouter benchmark exports63%
45DeepSeek V4 Flash 0731OpenRouter harness, deepseek/deepseek-v4-flash-0731, exa, server-tool, 25, high· OpenRouter benchmark exports60%
46GPT-5.6 LunaOpenRouter harness, openai/gpt-5.6-luna, perplexity, server-tool, 5, xhigh· OpenRouter benchmark exports60%
47GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, perplexity, server-tool, 5, high· OpenRouter benchmark exports59.8%
48GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, parallel, server-tool, 5, high· OpenRouter benchmark exports59.4%
49Claude Opus 5OpenRouter harness, anthropic/claude-opus-5, parallel, server-tool, 5, high· OpenRouter benchmark exports58%
50GPT-5.6 SolOpenRouter harness, openai/gpt-5.6-sol, exa, server-tool, 5, high· OpenRouter benchmark exports57.9%

Showing 1–50 of 67 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
Muse Spark (version unspecified)Meta Model API · highest effort · version unspecified · browser search/open/find · gpt-oss-120b answer extraction· Reported by Meta· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.76.8%

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. · OpenRouter public benchmark leaderboard and search-lane exports, CC BY 4.0. Fractional source scores converted to percentages; immutable dated snapshots. (licence)