| F1 score85.3%Reported 85.3% · Unit % Published Unconfirmed | DeepSearchQA · Low · compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Low effort;50K context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- low
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Low effort;3M total-token limit;50K context compaction;PTC enabled。
| View this score in rankings |
|---|
| F1 score87.9%Reported 87.9% · Unit % Published Unconfirmed | DeepSearchQA · Medium · compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Medium effort;50K context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- medium
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Medium effort;3M total-token limit;50K context compaction;PTC enabled。
| View this score in rankings |
|---|
| F1 score90.3%Reported 90.3% · Unit % Published Unconfirmed | DeepSearchQA · High · compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;High effort;50K context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- high
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;High effort;3M total-token limit;50K context compaction;PTC enabled。
| View this score in rankings |
|---|
| F1 score90.5%Reported 90.5% · Unit % Published Unconfirmed | DeepSearchQA · Max · compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Max effort;50K context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Max effort;3M total-token limit;50K context compaction;PTC enabled。
| View this score in rankings |
|---|
| F1 score84.5%Reported 84.5% · Unit % Published Unconfirmed | DeepSearchQA · Low · compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Low effort;50K context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- low
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Low effort;3M total-token limit;50K context compaction;PTC disabled。
| |
|---|
| F1 score87.3%Reported 87.3% · Unit % Published Unconfirmed | DeepSearchQA · Medium · compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Medium effort;50K context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- medium
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Medium effort;3M total-token limit;50K context compaction;PTC disabled。
| |
|---|
| F1 score87.8%Reported 87.8% · Unit % Published Unconfirmed | DeepSearchQA · High · compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;High effort;50K context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- high
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;High effort;3M total-token limit;50K context compaction;PTC disabled。
| |
|---|
| F1 score88.2%Reported 88.2% · Unit % Published Unconfirmed | DeepSearchQA · Max · compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Max effort;50K context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Max effort;3M total-token limit;50K context compaction;PTC disabled。
| |
|---|
| F1 score84.4%Reported 84.4% · Unit % Published Unconfirmed | DeepSearchQA · Low · no-compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Low effort;no context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- low
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Low effort;3M total-token limit;no context compaction;PTC enabled。
| |
|---|
| F1 score87.6%Reported 87.6% · Unit % Published Unconfirmed | DeepSearchQA · Medium · no-compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Medium effort;no context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- medium
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Medium effort;3M total-token limit;no context compaction;PTC enabled。
| |
|---|
| F1 score87.3%Reported 87.3% · Unit % Published Unconfirmed | DeepSearchQA · High · no-compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;High effort;no context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- high
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;High effort;3M total-token limit;no context compaction;PTC enabled。
| |
|---|
| F1 score87.7%Reported 87.7% · Unit % Published Unconfirmed | DeepSearchQA · Max · no-compaction-ptc · 3M- Settings
- Single agent;adaptive thinking;Max effort;no context compaction;PTC enabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Max effort;3M total-token limit;no context compaction;PTC enabled。
| |
|---|
| F1 score81.1%Reported 81.1% · Unit % Published Unconfirmed | DeepSearchQA · Low · no-compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Low effort;no context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- low
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Low effort;3M total-token limit;no context compaction;PTC disabled。
| |
|---|
| F1 score81.9%Reported 81.9% · Unit % Published Unconfirmed | DeepSearchQA · Medium · no-compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Medium effort;no context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- medium
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Medium effort;3M total-token limit;no context compaction;PTC disabled。
| |
|---|
| F1 score82.1%Reported 82.1% · Unit % Published Unconfirmed | DeepSearchQA · High · no-compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;High effort;no context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- high
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;High effort;3M total-token limit;no context compaction;PTC disabled。
| |
|---|
| F1 score82%Reported 82.0% · Unit % Published Unconfirmed | DeepSearchQA · Max · no-compaction-no-ptc · 3M- Settings
- Single agent;adaptive thinking;Max effort;no context compaction;PTC disabled;3M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tool access not recorded
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;no-compaction-no-ptc。
- Conditions
- Figure 2.21.3.1.A; one run per datapoint; possible noise;adaptive thinking enabled;Max effort;3M total-token limit;no context compaction;PTC disabled。
| |
|---|
| F1 score91.3%Reported 91.3% · Unit % Published Unconfirmed | DeepSearchQA · Max · 50K compaction + PTC · 10M- Settings
- Single agent;adaptive thinking;Max effort;50K context compaction;PTC;10M total-token limit。
- Unknowns
- Figure 2.21.3.1.A reports one run per datapoint; the card notes possible noise.
- Tools
- Tools allowed
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;single-agent;50K context compaction;PTC。
- Conditions
- 10M total-token limit;adaptive thinking、max effort;standard single-agent configuration;also reported in Figure 2.21.3.A.。
| View this score in rankings |
|---|
| F1 score92.5%Reported 92.5% · Unit % Published Unconfirmed | DeepSearchQA · multi-agent · Max- Settings
- Orchestrator 无直接工具,仅委派;orchestrator/subagents 均以 50K 触发 compaction、各自最多 3M tokens;web search/fetch、PTC、Max effort。
- Unknowns
- 多代理结果与单代理成绩分列。
- Tools
- Tools allowed
- Reasoning budget
- max
- Framework
- not recorded
- Protocol
- DeepSearchQA;multi-agent;orchestrator and subagents。
- Conditions
- 50K context compaction;up to 3M tokens for each orchestrator/subagent; web search/fetch and PTC; max effort. Kept separate from single-agent scores.。
| |
|---|
| Accuracy84%Reported 84.0% · Unit % Published Unconfirmed | DeepSearchQA · main figure accuracy- Settings
- The main figure labels the Opus 4.6 value as accuracy.
- Unknowns
- The figure does not specify agent setup, token budget, reasoning effort or trial count; kept separate from card F1 results.
- Tools
- Tool access not recorded
- Reasoning budget
- not recorded
- Framework
- not recorded
- Protocol
- DeepSearchQA; accuracy as labeled by the main figure.
- Conditions
- Agent setup, token budget, effort and trial count are not stated. Kept separate from the System Card's DeepSearchQA F1 configurations.
| View this score in rankings |
|---|