Agentic AI Advanced MCQ Questions with Answers (Latest 2026)

Q1. Which option best describes Reflexion in agentic AI?

A. A retrieval method over reflections. B. An RL algorithm tuning model weights online. C. Self-critique loop where the agent reflects on failures and revises strategy. D. A speculative decoding technique.

Select an answer to check.

Answer: Self-critique loop where the agent reflects on failures and revises strategy.

Here, Self-critique loop where the agent reflects on failures and revises strategy. is the right choice. Reflexion improves performance via written self-feedback. It aligns directly with what the question asks about which option best describes reflexion in agentic ai. A quick elimination of partially true options helps confirm it.

Q2. What is the primary purpose of Reflexion?

A. A speculative decoding technique. B. Self-critique loop where the agent reflects on failures and revises strategy. C. A retrieval method over reflections. D. An RL algorithm tuning model weights online.

Select an answer to check.

Answer: Self-critique loop where the agent reflects on failures and revises strategy.

In this case, Self-critique loop where the agent reflects on failures and revises strategy. is correct. Reflexion improves performance via written self-feedback. It aligns directly with what the question asks about what is the primary purpose of reflexion. A quick elimination of partially true options helps confirm it.

Q3. Which statement about Reflexion is most accurate?

A. An RL algorithm tuning model weights online. B. Self-critique loop where the agent reflects on failures and revises strategy. C. A retrieval method over reflections. D. A speculative decoding technique.

Select an answer to check.

Answer: Self-critique loop where the agent reflects on failures and revises strategy.

The best option here is Self-critique loop where the agent reflects on failures and revises strategy.. Reflexion improves performance via written self-feedback. It aligns directly with what the question asks about which statement about reflexion is most accurate. A quick elimination of partially true options helps confirm it.

Q4. How is Reflexion best characterized?

A. A retrieval method over reflections. B. Self-critique loop where the agent reflects on failures and revises strategy. C. A speculative decoding technique. D. An RL algorithm tuning model weights online.

Select an answer to check.

Answer: Self-critique loop where the agent reflects on failures and revises strategy.

For this question, Self-critique loop where the agent reflects on failures and revises strategy. is correct. Reflexion improves performance via written self-feedback. It aligns directly with what the question asks about how is reflexion best characterized. A quick elimination of partially true options helps confirm it.

Q5. Which option best describes Tree-of-Thoughts in agentic AI?

A. A linear chain-of-thought trace. B. A graph DB query plan. C. Search over multiple reasoning branches with evaluation and pruning. D. A static decision tree classifier.

Select an answer to check.

Answer: Search over multiple reasoning branches with evaluation and pruning.

Search over multiple reasoning branches with evaluation and pruning. is the correct answer here. ToT explores alternative reasoning paths. It aligns directly with what the question asks about which option best describes tree-of-thoughts in agentic ai. A quick elimination of partially true options helps confirm it.

Q6. What is the primary purpose of Tree-of-Thoughts?

A. A graph DB query plan. B. Search over multiple reasoning branches with evaluation and pruning. C. A linear chain-of-thought trace. D. A static decision tree classifier.

Select an answer to check.

Answer: Search over multiple reasoning branches with evaluation and pruning.

Here, Search over multiple reasoning branches with evaluation and pruning. is the right choice. ToT explores alternative reasoning paths. This matches the core idea being tested around what is the primary purpose of tree-of-thoughts. A quick elimination of partially true options helps confirm it.

Q7. Which statement about Tree-of-Thoughts is most accurate?

A. Search over multiple reasoning branches with evaluation and pruning. B. A graph DB query plan. C. A linear chain-of-thought trace. D. A static decision tree classifier.

Select an answer to check.

Answer: Search over multiple reasoning branches with evaluation and pruning.

In this case, Search over multiple reasoning branches with evaluation and pruning. is correct. ToT explores alternative reasoning paths. This matches the core idea being tested around which statement about tree-of-thoughts is most accurate. A quick elimination of partially true options helps confirm it.

Q8. How is Tree-of-Thoughts best characterized?

A. Search over multiple reasoning branches with evaluation and pruning. B. A linear chain-of-thought trace. C. A static decision tree classifier. D. A graph DB query plan.

Select an answer to check.

Answer: Search over multiple reasoning branches with evaluation and pruning.

The best option here is Search over multiple reasoning branches with evaluation and pruning.. ToT explores alternative reasoning paths. This matches the core idea being tested around how is tree-of-thoughts best characterized. A quick elimination of partially true options helps confirm it.

Q9. Which option best describes Plan-and-Execute in agentic AI?

A. An RL training schedule. B. A vector retrieval ranker. C. Two-phase pattern: produce a full plan, then execute steps with a worker. D. A single-step tool call.

Select an answer to check.

Answer: Two-phase pattern: produce a full plan, then execute steps with a worker.

For this question, Two-phase pattern: produce a full plan, then execute steps with a worker. is correct. Decouples planning cost from per-step execution. This matches the core idea being tested around which option best describes plan-and-execute in agentic ai. A quick elimination of partially true options helps confirm it.

Q10. What is the primary purpose of Plan-and-Execute?

A. Two-phase pattern: produce a full plan, then execute steps with a worker. B. An RL training schedule. C. A vector retrieval ranker. D. A single-step tool call.

Select an answer to check.

Answer: Two-phase pattern: produce a full plan, then execute steps with a worker.

Two-phase pattern: produce a full plan, then execute steps with a worker. is the correct answer here. Decouples planning cost from per-step execution. This matches the core idea being tested around what is the primary purpose of plan-and-execute. A quick elimination of partially true options helps confirm it.

Q11. Which statement about Plan-and-Execute is most accurate?

A. A single-step tool call. B. A vector retrieval ranker. C. Two-phase pattern: produce a full plan, then execute steps with a worker. D. An RL training schedule.

Select an answer to check.

Answer: Two-phase pattern: produce a full plan, then execute steps with a worker.

Here, Two-phase pattern: produce a full plan, then execute steps with a worker. is the right choice. Decouples planning cost from per-step execution. That is exactly the concept behind which statement about plan-and-execute is most accurate in this context. A quick elimination of partially true options helps confirm it.

Q12. How is Plan-and-Execute best characterized?

A. An RL training schedule. B. A vector retrieval ranker. C. Two-phase pattern: produce a full plan, then execute steps with a worker. D. A single-step tool call.

Select an answer to check.

Answer: Two-phase pattern: produce a full plan, then execute steps with a worker.

In this case, Two-phase pattern: produce a full plan, then execute steps with a worker. is correct. Decouples planning cost from per-step execution. That is exactly the concept behind how is plan-and-execute best characterized in this context. A quick elimination of partially true options helps confirm it.

Q13. Which option best describes self-consistency decoding in agentic AI?

A. A beam search over tools. B. Sample multiple reasoning paths and pick the majority answer. C. A token-level constrained decoder. D. Greedy decoding with temperature 0.

Select an answer to check.

Answer: Sample multiple reasoning paths and pick the majority answer.

The best option here is Sample multiple reasoning paths and pick the majority answer.. Improves robustness on reasoning tasks. That is exactly the concept behind which option best describes self-consistency decoding in agentic in this context. A quick elimination of partially true options helps confirm it.

Q14. What is the primary purpose of self-consistency decoding?

A. Sample multiple reasoning paths and pick the majority answer. B. A token-level constrained decoder. C. Greedy decoding with temperature 0. D. A beam search over tools.

Select an answer to check.

Answer: Sample multiple reasoning paths and pick the majority answer.

For this question, Sample multiple reasoning paths and pick the majority answer. is correct. Improves robustness on reasoning tasks. That is exactly the concept behind what is the primary purpose of self-consistency decoding in this context. A quick elimination of partially true options helps confirm it.

Q15. Which statement about self-consistency decoding is most accurate?

A. Greedy decoding with temperature 0. B. A beam search over tools. C. A token-level constrained decoder. D. Sample multiple reasoning paths and pick the majority answer.

Select an answer to check.

Answer: Sample multiple reasoning paths and pick the majority answer.

Sample multiple reasoning paths and pick the majority answer. is the correct answer here. Improves robustness on reasoning tasks. That is exactly the concept behind which statement about self-consistency decoding is most accurate in this context. A quick elimination of partially true options helps confirm it.

Q16. How is self-consistency decoding best characterized?

A. A token-level constrained decoder. B. Greedy decoding with temperature 0. C. A beam search over tools. D. Sample multiple reasoning paths and pick the majority answer.

Select an answer to check.

Answer: Sample multiple reasoning paths and pick the majority answer.

Here, Sample multiple reasoning paths and pick the majority answer. is the right choice. Improves robustness on reasoning tasks. It fits the requirement in the prompt about how is self-consistency decoding best characterized. A quick elimination of partially true options helps confirm it.

Q17. Which option best describes LangGraph in agentic AI?

A. A model fine-tuning library. B. A graph-based agent orchestration framework with explicit nodes/edges. C. A prompt-template DSL. D. A vector database.

Select an answer to check.

Answer: A graph-based agent orchestration framework with explicit nodes/edges.

In this case, A graph-based agent orchestration framework with explicit nodes/edges. is correct. LangGraph models agent control flow as a graph. It fits the requirement in the prompt about which option best describes langgraph in agentic ai. A quick elimination of partially true options helps confirm it.

Q18. What is the primary purpose of LangGraph?

A. A vector database. B. A model fine-tuning library. C. A graph-based agent orchestration framework with explicit nodes/edges. D. A prompt-template DSL.

Select an answer to check.

Answer: A graph-based agent orchestration framework with explicit nodes/edges.

The best option here is A graph-based agent orchestration framework with explicit nodes/edges.. LangGraph models agent control flow as a graph. It fits the requirement in the prompt about what is the primary purpose of langgraph. A quick elimination of partially true options helps confirm it.

Q19. Which statement about LangGraph is most accurate?

A. A prompt-template DSL. B. A model fine-tuning library. C. A vector database. D. A graph-based agent orchestration framework with explicit nodes/edges.

Select an answer to check.

Answer: A graph-based agent orchestration framework with explicit nodes/edges.

For this question, A graph-based agent orchestration framework with explicit nodes/edges. is correct. LangGraph models agent control flow as a graph. It fits the requirement in the prompt about which statement about langgraph is most accurate. A quick elimination of partially true options helps confirm it.

Q20. How is LangGraph best characterized?

A. A model fine-tuning library. B. A prompt-template DSL. C. A vector database. D. A graph-based agent orchestration framework with explicit nodes/edges.

Select an answer to check.

Answer: A graph-based agent orchestration framework with explicit nodes/edges.

A graph-based agent orchestration framework with explicit nodes/edges. is the correct answer here. LangGraph models agent control flow as a graph. It fits the requirement in the prompt about how is langgraph best characterized. A quick elimination of partially true options helps confirm it.

Q21. Which option best describes AutoGen in agentic AI?

A. A retrieval index format. B. A code formatter. C. A dataset generator. D. A multi-agent conversation framework from Microsoft.

Select an answer to check.

Answer: A multi-agent conversation framework from Microsoft.

Here, A multi-agent conversation framework from Microsoft. is the right choice. AutoGen orchestrates conversing agents and tools. This is the most accurate statement for which option best describes autogen in agentic ai. A quick elimination of partially true options helps confirm it.

Q22. What is the primary purpose of AutoGen?

A. A code formatter. B. A dataset generator. C. A retrieval index format. D. A multi-agent conversation framework from Microsoft.

Select an answer to check.

Answer: A multi-agent conversation framework from Microsoft.

In this case, A multi-agent conversation framework from Microsoft. is correct. AutoGen orchestrates conversing agents and tools. This is the most accurate statement for what is the primary purpose of autogen. A quick elimination of partially true options helps confirm it.

Q23. Which statement about AutoGen is most accurate?

A. A code formatter. B. A multi-agent conversation framework from Microsoft. C. A retrieval index format. D. A dataset generator.

Select an answer to check.

Answer: A multi-agent conversation framework from Microsoft.

The best option here is A multi-agent conversation framework from Microsoft.. AutoGen orchestrates conversing agents and tools. This is the most accurate statement for which statement about autogen is most accurate. A quick elimination of partially true options helps confirm it.

Q24. How is AutoGen best characterized?

A. A dataset generator. B. A multi-agent conversation framework from Microsoft. C. A retrieval index format. D. A code formatter.

Select an answer to check.

Answer: A multi-agent conversation framework from Microsoft.

For this question, A multi-agent conversation framework from Microsoft. is correct. AutoGen orchestrates conversing agents and tools. This is the most accurate statement for how is autogen best characterized. A quick elimination of partially true options helps confirm it.

Q25. Which option best describes a supervisor agent in agentic AI?

A. A tokenizer wrapper. B. A retrieval reranker. C. A coordinator that delegates work to specialized sub-agents. D. A guardrail validator.

Select an answer to check.

Answer: A coordinator that delegates work to specialized sub-agents.

A coordinator that delegates work to specialized sub-agents. is the correct answer here. Supervisors enable hierarchical agent structures. This is the most accurate statement for which option best describes a supervisor agent in. A quick elimination of partially true options helps confirm it.

Q26. What is the primary purpose of a supervisor agent?

A. A tokenizer wrapper. B. A coordinator that delegates work to specialized sub-agents. C. A guardrail validator. D. A retrieval reranker.

Select an answer to check.

Answer: A coordinator that delegates work to specialized sub-agents.

Here, A coordinator that delegates work to specialized sub-agents. is the right choice. Supervisors enable hierarchical agent structures. It aligns directly with what the question asks about what is the primary purpose of a supervisor. The other options are either incomplete or contextually incorrect.

Q27. Which statement about a supervisor agent is most accurate?

A. A guardrail validator. B. A coordinator that delegates work to specialized sub-agents. C. A retrieval reranker. D. A tokenizer wrapper.

Select an answer to check.

Answer: A coordinator that delegates work to specialized sub-agents.

In this case, A coordinator that delegates work to specialized sub-agents. is correct. Supervisors enable hierarchical agent structures. It aligns directly with what the question asks about which statement about a supervisor agent is most. The other options are either incomplete or contextually incorrect.

Q28. How is a supervisor agent best characterized?

A. A coordinator that delegates work to specialized sub-agents. B. A guardrail validator. C. A retrieval reranker. D. A tokenizer wrapper.

Select an answer to check.

Answer: A coordinator that delegates work to specialized sub-agents.

The best option here is A coordinator that delegates work to specialized sub-agents.. Supervisors enable hierarchical agent structures. It aligns directly with what the question asks about how is a supervisor agent best characterized. The other options are either incomplete or contextually incorrect.

Q29. Which option best describes chain-of-verification in agentic AI?

A. A speculative decoding pass. B. A pre-training data filter. C. A tokenizer step. D. Generating verification questions to check a draft answer.

Select an answer to check.

Answer: Generating verification questions to check a draft answer.

For this question, Generating verification questions to check a draft answer. is correct. CoVe reduces hallucinations through self-checks. It aligns directly with what the question asks about which option best describes chain-of-verification in agentic ai. The other options are either incomplete or contextually incorrect.

Q30. What is the primary purpose of chain-of-verification?

A. A speculative decoding pass. B. A pre-training data filter. C. Generating verification questions to check a draft answer. D. A tokenizer step.

Select an answer to check.

Answer: Generating verification questions to check a draft answer.

Generating verification questions to check a draft answer. is the correct answer here. CoVe reduces hallucinations through self-checks. It aligns directly with what the question asks about what is the primary purpose of chain-of-verification. The other options are either incomplete or contextually incorrect.

Q31. Which statement about chain-of-verification is most accurate?

A. A speculative decoding pass. B. A pre-training data filter. C. A tokenizer step. D. Generating verification questions to check a draft answer.

Select an answer to check.

Answer: Generating verification questions to check a draft answer.

Here, Generating verification questions to check a draft answer. is the right choice. CoVe reduces hallucinations through self-checks. This matches the core idea being tested around which statement about chain-of-verification is most accurate. The other options are either incomplete or contextually incorrect.

Q32. How is chain-of-verification best characterized?

A. Generating verification questions to check a draft answer. B. A pre-training data filter. C. A tokenizer step. D. A speculative decoding pass.

Select an answer to check.

Answer: Generating verification questions to check a draft answer.

In this case, Generating verification questions to check a draft answer. is correct. CoVe reduces hallucinations through self-checks. This matches the core idea being tested around how is chain-of-verification best characterized. The other options are either incomplete or contextually incorrect.

Q33. Which option best describes speculative decoding in agentic AI?

A. A vector store sharding strategy. B. A multi-agent debate protocol. C. A retrieval reranker. D. Drafting tokens with a small model and verifying with a large one.

Select an answer to check.

Answer: Drafting tokens with a small model and verifying with a large one.

The best option here is Drafting tokens with a small model and verifying with a large one.. Speeds inference without changing output distribution. This matches the core idea being tested around which option best describes speculative decoding in agentic. The other options are either incomplete or contextually incorrect.

Q34. What is the primary purpose of speculative decoding?

A. A vector store sharding strategy. B. A multi-agent debate protocol. C. A retrieval reranker. D. Drafting tokens with a small model and verifying with a large one.

Select an answer to check.

Answer: Drafting tokens with a small model and verifying with a large one.

For this question, Drafting tokens with a small model and verifying with a large one. is correct. Speeds inference without changing output distribution. This matches the core idea being tested around what is the primary purpose of speculative decoding. The other options are either incomplete or contextually incorrect.

Q35. Which statement about speculative decoding is most accurate?

A. A multi-agent debate protocol. B. Drafting tokens with a small model and verifying with a large one. C. A retrieval reranker. D. A vector store sharding strategy.

Select an answer to check.

Answer: Drafting tokens with a small model and verifying with a large one.

Drafting tokens with a small model and verifying with a large one. is the correct answer here. Speeds inference without changing output distribution. This matches the core idea being tested around which statement about speculative decoding is most accurate. The other options are either incomplete or contextually incorrect.

Q36. How is speculative decoding best characterized?

A. Drafting tokens with a small model and verifying with a large one. B. A vector store sharding strategy. C. A retrieval reranker. D. A multi-agent debate protocol.

Select an answer to check.

Answer: Drafting tokens with a small model and verifying with a large one.

Here, Drafting tokens with a small model and verifying with a large one. is the right choice. Speeds inference without changing output distribution. That is exactly the concept behind how is speculative decoding best characterized in this context. The other options are either incomplete or contextually incorrect.

Q37. Which option best describes constrained decoding in agentic AI?

A. A reinforcement-learning fine-tune. B. A cache eviction policy. C. Restricting generated tokens to a grammar or schema. D. A retrieval filter.

Select an answer to check.

Answer: Restricting generated tokens to a grammar or schema.

In this case, Restricting generated tokens to a grammar or schema. is correct. Guarantees structured output formats. That is exactly the concept behind which option best describes constrained decoding in agentic in this context. The other options are either incomplete or contextually incorrect.

Q38. What is the primary purpose of constrained decoding?

A. Restricting generated tokens to a grammar or schema. B. A retrieval filter. C. A reinforcement-learning fine-tune. D. A cache eviction policy.

Select an answer to check.

Answer: Restricting generated tokens to a grammar or schema.

The best option here is Restricting generated tokens to a grammar or schema.. Guarantees structured output formats. That is exactly the concept behind what is the primary purpose of constrained decoding in this context. The other options are either incomplete or contextually incorrect.

Q39. Which statement about constrained decoding is most accurate?

A. A cache eviction policy. B. A reinforcement-learning fine-tune. C. A retrieval filter. D. Restricting generated tokens to a grammar or schema.

Select an answer to check.

Answer: Restricting generated tokens to a grammar or schema.

For this question, Restricting generated tokens to a grammar or schema. is correct. Guarantees structured output formats. That is exactly the concept behind which statement about constrained decoding is most accurate in this context. The other options are either incomplete or contextually incorrect.

Q40. How is constrained decoding best characterized?

A. Restricting generated tokens to a grammar or schema. B. A reinforcement-learning fine-tune. C. A cache eviction policy. D. A retrieval filter.

Select an answer to check.

Answer: Restricting generated tokens to a grammar or schema.

Restricting generated tokens to a grammar or schema. is the correct answer here. Guarantees structured output formats. That is exactly the concept behind how is constrained decoding best characterized in this context. The other options are either incomplete or contextually incorrect.

Q41. Which option best describes structured tool use in agentic AI?

A. A heuristic command dispatcher. B. A regex-based parser of model output. C. Free-form text instructions to a tool. D. Tool invocation via validated JSON schemas with typed args.

Select an answer to check.

Answer: Tool invocation via validated JSON schemas with typed args.

Here, Tool invocation via validated JSON schemas with typed args. is the right choice. Schemas avoid brittle string parsing. It fits the requirement in the prompt about which option best describes structured tool use in. The other options are either incomplete or contextually incorrect.

Q42. What is the primary purpose of structured tool use?

A. A heuristic command dispatcher. B. Tool invocation via validated JSON schemas with typed args. C. Free-form text instructions to a tool. D. A regex-based parser of model output.

Select an answer to check.

Answer: Tool invocation via validated JSON schemas with typed args.

In this case, Tool invocation via validated JSON schemas with typed args. is correct. Schemas avoid brittle string parsing. It fits the requirement in the prompt about what is the primary purpose of structured tool. The other options are either incomplete or contextually incorrect.

Q43. Which statement about structured tool use is most accurate?

A. A heuristic command dispatcher. B. A regex-based parser of model output. C. Free-form text instructions to a tool. D. Tool invocation via validated JSON schemas with typed args.

Select an answer to check.

Answer: Tool invocation via validated JSON schemas with typed args.

The best option here is Tool invocation via validated JSON schemas with typed args.. Schemas avoid brittle string parsing. It fits the requirement in the prompt about which statement about structured tool use is most. The other options are either incomplete or contextually incorrect.

Q44. How is structured tool use best characterized?

A. Tool invocation via validated JSON schemas with typed args. B. Free-form text instructions to a tool. C. A regex-based parser of model output. D. A heuristic command dispatcher.

Select an answer to check.

Answer: Tool invocation via validated JSON schemas with typed args.

For this question, Tool invocation via validated JSON schemas with typed args. is correct. Schemas avoid brittle string parsing. It fits the requirement in the prompt about how is structured tool use best characterized. The other options are either incomplete or contextually incorrect.

Q45. Which option best describes agent-as-a-tool in agentic AI?

A. Embedding agent logs. B. Wrapping a sub-agent so a parent agent can call it like any tool. C. Caching a prompt template. D. Storing an agent's weights as a file.

Select an answer to check.

Answer: Wrapping a sub-agent so a parent agent can call it like any tool.

Wrapping a sub-agent so a parent agent can call it like any tool. is the correct answer here. Enables composition of agents. It fits the requirement in the prompt about which option best describes agent-as-a-tool in agentic ai. The other options are either incomplete or contextually incorrect.

Q46. What is the primary purpose of agent-as-a-tool?

A. Embedding agent logs. B. Wrapping a sub-agent so a parent agent can call it like any tool. C. Storing an agent's weights as a file. D. Caching a prompt template.

Select an answer to check.

Answer: Wrapping a sub-agent so a parent agent can call it like any tool.

Here, Wrapping a sub-agent so a parent agent can call it like any tool. is the right choice. Enables composition of agents. This is the most accurate statement for what is the primary purpose of agent-as-a-tool. The other options are either incomplete or contextually incorrect.

Q47. Which statement about agent-as-a-tool is most accurate?

A. Embedding agent logs. B. Caching a prompt template. C. Wrapping a sub-agent so a parent agent can call it like any tool. D. Storing an agent's weights as a file.

Select an answer to check.

Answer: Wrapping a sub-agent so a parent agent can call it like any tool.

In this case, Wrapping a sub-agent so a parent agent can call it like any tool. is correct. Enables composition of agents. This is the most accurate statement for which statement about agent-as-a-tool is most accurate. The other options are either incomplete or contextually incorrect.

Q48. How is agent-as-a-tool best characterized?

A. Caching a prompt template. B. Embedding agent logs. C. Wrapping a sub-agent so a parent agent can call it like any tool. D. Storing an agent's weights as a file.

Select an answer to check.

Answer: Wrapping a sub-agent so a parent agent can call it like any tool.

The best option here is Wrapping a sub-agent so a parent agent can call it like any tool.. Enables composition of agents. This is the most accurate statement for how is agent-as-a-tool best characterized. The other options are either incomplete or contextually incorrect.

Q49. Which option best describes token budget management in agentic AI?

A. Selecting a smaller model only. B. Lowering temperature. C. Disabling logging. D. Tracking and bounding tokens spent across an agent run.

Select an answer to check.

Answer: Tracking and bounding tokens spent across an agent run.

For this question, Tracking and bounding tokens spent across an agent run. is correct. Prevents runaway cost in long loops. This is the most accurate statement for which option best describes token budget management in. The other options are either incomplete or contextually incorrect.

Q50. What is the primary purpose of token budget management?

A. Tracking and bounding tokens spent across an agent run. B. Lowering temperature. C. Disabling logging. D. Selecting a smaller model only.

Select an answer to check.

Answer: Tracking and bounding tokens spent across an agent run.

Tracking and bounding tokens spent across an agent run. is the correct answer here. Prevents runaway cost in long loops. This is the most accurate statement for what is the primary purpose of token budget. The other options are either incomplete or contextually incorrect.