Research & Academia
Whole subtree ・ 0 votes
This is the combined evaluation of votes for AI Search & Deep Research and Math.
A hub comparing AI tools for research — from AI search and deep research to math problem-solving.
Sub-categoriesVoting happens on each sub-category page
No votes yet. Be the first voter.
Rating criteria
※ Your tier is simply your own subjective judgment — that's fine. The following are rough guidelines when voting.
- SRecommendedExecutes tasks effortlessly.Produces the intended output with few prompts and turns.Excellent cost performance.Almost no hallucinations.
- AGoodExperienced users can accomplish most tasks.Eventually produces the intended output.Good cost performance.Few hallucinations.
- BAverageProduces the intended output only after many prompts and turns.Average cost performance.Errors and hallucinations occur.
- CSomewhat flawedFairly likely to never produce the intended output no matter how many turns you take.Frequent errors and hallucinations.
- ✗Nearly impossibleFails to produce the intended output in most cases.
Vote from a sub-category
This tier list aggregates votes cast in the sub-categories. Pick the one you want to vote in.
External benchmark composite (reference)
Papers, review sites, and hands-on media tests converted to tiers and averaged (as of 2026-10-15)
Simple average of per-source tiers (S=4 … C=1). Treat single-source tiers with caution (source count in parentheses).
Averaged from the subcategory benchmark tables and Arena tier tables, with tiers scored S=4 to C=1 (AI Search & Deep Research (Arena), Math (Arena)).
Sources:Arena (search)Arena (math)