Search & research

• Best overallSGPT-6.1 Sol1,250 on LMArena Search Arena (style control)
• Best valueAGPT-6 LunaTier A for $0.2/M tokens
• Best free startSExaFree tier · paid plans · 80 on Artificial Analysis Search Index (Search APIs)
• Best local (≤24 GB GPU)No qualifying option yet

• SABCTier: how close to the best result on the task's main public benchmark
• A Outlined: a tool inherits the tier of the best model it offers
• – No public result yet
• Cloud · Local: runs on your own hardware

CloudLocal

Models

The engines themselves. Use them through an API or inside the apps on the right.

Cloud

S GPT-6.1 SolOpenAI · $4/M tokens1,250 on LMArena Search Arena (style control)$S GPT-6 AstraOpenAI · $20/M tokens1,246 on LMArena Search Arena (style control)S Claude Fable 5.1Anthropic · $20/M tokens1,233 on LMArena Search Arena (style control)S Qwen3.8-MaxAlibaba (Qwen) · $3/M tokens1,230 on LMArena Search Arena (style control)$A Claude Opus 5.5Anthropic · $8/M tokens1,211 on LMArena Search Arena (style control)A GPT-6 LunaOpenAI · $0.2/M tokens1,209 on LMArena Search Arena (style control)$A Gemini 3.1 Pro (Preview)Google · $4.5/M tokens1,207 on LMArena Search Arena (style control)A ERNIE 5.1Baidu1,192 on LMArena Search Arena (style control)B Sonar ProPerplexity AI · $6/M tokens1,143 on LMArena Search Arena (style control)B Sonar Reasoning ProPerplexity AI · $3.5/M tokens1,139 on LMArena Search Arena (style control)– Claude Sonnet 5.5Anthropic · $4/M tokensNo public result yet– Gemini 3.8 FlashGoogle · $1.5/M tokensNo public result yet– Grok 4.7xAI · $3/M tokensNo public result yet– SonarPerplexity AI · $1/M tokensNo public result yet– Sonar Deep ResearchPerplexity AI · $3.5/M tokensNo public result yet

Local

Nothing tracked yet

Apps

Ready-made apps with a UI: sign in or install, then start.

Cloud

S SciraScira · Free tier · $15/movia GPT-6.1 SolS Microsoft CopilotMicrosoft · Free tier · $9.99/movia GPT-6.1 Sol$S ClaudeAnthropic · Free tier · $20/movia Claude Fable 5.1A GeminiGoogle · Free tier · $4.99/mo50 on DeepResearch Bench (RACE, GPT-5.5 judge)$A ChatGPTOpenAI · Free tier · $20/mo48 on DeepResearch Bench (RACE, GPT-5.5 judge)A Kagi AssistantKagi · Free tier · $5/movia Gemini 3.1 Pro (Preview)$B PerplexityPerplexity AI · Free tier · $20/mo43 on DeepResearch Bench (RACE, GPT-5.5 judge)B GrokSpaceXSI (formerly xAI) · Free tier · $8/mo41 on DeepResearch Bench (RACE, GPT-5.5 judge)– Brave SearchBraveNo public result yet– ConsensusConsensusNo public result yet– DeepSeekDeepSeek · FreeNo public result yet– DoubaoByteDance · Free tierNo public result yet– ElicitElicit · Free tier · $49/moNo public result yet– FeloFelo · Free tier · $15/moNo public result yet– Google AI ModeGoogle · FreeNo public result yet– iAskiAsk.Ai · Free tierNo public result yet– KimiMoonshot AI · Free tierNo public result yet– KomoKomo · Free tierNo public result yet– LinerLiner · Free tier · $18/moNo public result yet– Meta AIMeta · Free tierNo public result yet– Metaso AI SearchMetaso (秘塔科技)No public result yet– Qwen ChatAlibaba (Qwen team) · FreeNo public result yet– UndermindUndermind · Free tier · $16/moNo public result yet– Vibe (formerly Le Chat)Mistral AI · Free tier · $15/moNo public result yet

Local

– Open WebUIOpen WebUI Inc. · FreeNo public result yet– KhojKhoj · FreeNo public result yet– Vane (formerly Perplexica)ItzCrazyKns (open-source project) · FreeNo public result yet– Local Deep ResearchLearningCircuit (open-source project) · FreeNo public result yet– MorphicMorphic (miurla) · FreeNo public result yet

Search APIs & engines

Web search for your own apps and agents, and open search or research engines you can host yourself.

Cloud

S ExaExa · Free tier80 on Artificial Analysis Search Index (Search APIs)S Perplexity Search APIPerplexity AI80 on Artificial Analysis Search Index (Search APIs)S OpenRouterOpenRouter · Free tiervia GPT-6.1 SolA ParallelParallel Web Systems · Free tier75 on Artificial Analysis Search Index (Search APIs)A Brave Search APIBrave · Free tier75 on Artificial Analysis Search Index (Search APIs)A You.com APIsYou.com · Free tier74 on Artificial Analysis Search Index (Search APIs)B LinkupLinkup · Free tier68 on Artificial Analysis Search Index (Search APIs)B PerplexityPerplexity AI · Free tier · $20/mo43 on DeepResearch Bench (RACE, GPT-5.5 judge)$B TavilyTavily · Free tier · $30/mo66 on Artificial Analysis Search Index (Search APIs)– Jina Search Foundation APIsJina AI (Elastic) · Free tierNo public result yet– ValyuValyuNo public result yet

Local

– FirecrawlFirecrawl · Free tier · $19/moNo public result yet– SearXNGSearXNG (open-source project) · FreeNo public result yet– GPT ResearcherGPT Researcher (open-source project) · FreeNo public result yet

Everything, in order

All search & research options in quality order

Tier	Name	Approach	Runs	Rated by	Price	Skill
S	Exa Exa	Search APIs & engines	Cloud	80 on Artificial Analysis Search Index (Search APIs)	Free tier	intermediate
S	Perplexity Search API Perplexity AI	Search APIs & engines	Cloud	80 on Artificial Analysis Search Index (Search APIs)	—	intermediate
S	GPT-6.1 Sol OpenAI	Models	Cloud	1,250 on LMArena Search Arena (style control)	$4/M tokens	—
S	GPT-6 Astra OpenAI	Models	Cloud	1,246 on LMArena Search Arena (style control)	$20/M tokens	—
S	Claude Fable 5.1 Anthropic	Models	Cloud	1,233 on LMArena Search Arena (style control)	$20/M tokens	—
S	Qwen3.8-Max Alibaba (Qwen)	Models	Cloud	1,230 on LMArena Search Arena (style control)	$3/M tokens	—
S	Scira Scira	Apps	Cloud	via GPT-6.1 Sol	Free tier · $15/mo	beginner
S	Microsoft Copilot Microsoft	Apps	Cloud	via GPT-6.1 Sol	Free tier · $9.99/mo	beginner
S	OpenRouter OpenRouter	Search APIs & engines	Cloud	via GPT-6.1 Sol	Free tier	intermediate
S	Claude Anthropic	Apps	Cloud	via Claude Fable 5.1	Free tier · $20/mo	beginner
A	Claude Opus 5.5 Anthropic	Models	Cloud	1,211 on LMArena Search Arena (style control)	$8/M tokens	—
A	GPT-6 Luna OpenAI	Models	Cloud	1,209 on LMArena Search Arena (style control)	$0.2/M tokens	—
A	Gemini 3.1 Pro (Preview) Google	Models	Cloud	1,207 on LMArena Search Arena (style control)	$4.5/M tokens	—
A	Parallel Parallel Web Systems	Search APIs & engines	Cloud	75 on Artificial Analysis Search Index (Search APIs)	Free tier	intermediate
A	Brave Search API Brave	Search APIs & engines	Cloud	75 on Artificial Analysis Search Index (Search APIs)	Free tier	intermediate
A	Gemini Google	Apps	Cloud	50 on DeepResearch Bench (RACE, GPT-5.5 judge)	Free tier · $4.99/mo	beginner
A	You.com APIs You.com	Search APIs & engines	Cloud	74 on Artificial Analysis Search Index (Search APIs)	Free tier	intermediate
A	ERNIE 5.1 Baidu	Models	Cloud	1,192 on LMArena Search Arena (style control)	—	—
A	ChatGPT OpenAI	Apps	Cloud	48 on DeepResearch Bench (RACE, GPT-5.5 judge)	Free tier · $20/mo	beginner
A	Kagi Assistant Kagi	Apps	Cloud	via Gemini 3.1 Pro (Preview)	Free tier · $5/mo	beginner
B	Sonar Pro Perplexity AI	Models	Cloud	1,143 on LMArena Search Arena (style control)	$6/M tokens	—
B	Sonar Reasoning Pro Perplexity AI	Models	Cloud	1,139 on LMArena Search Arena (style control)	$3.5/M tokens	—
B	Linkup Linkup	Search APIs & engines	Cloud	68 on Artificial Analysis Search Index (Search APIs)	Free tier	intermediate
B	Perplexity Perplexity AI	Apps + Search APIs & engines	Cloud	43 on DeepResearch Bench (RACE, GPT-5.5 judge)	Free tier · $20/mo	beginner
B	Grok SpaceXSI (formerly xAI)	Apps	Cloud	41 on DeepResearch Bench (RACE, GPT-5.5 judge)	Free tier · $8/mo	beginner
B	Tavily Tavily	Search APIs & engines	Cloud	66 on Artificial Analysis Search Index (Search APIs)	Free tier · $30/mo	intermediate
–	Firecrawl Firecrawl	Search APIs & engines	Local	No public result yet Cloud models: tier A	Free tier · $19/mo	intermediate
–	Open WebUI Open WebUI Inc.	Apps	Local	No public result yet	Free	intermediate
–	SearXNG SearXNG (open-source project)	Search APIs & engines	Local	No public result yet	Free	advanced
–	Khoj Khoj	Apps	Local	No public result yet	Free	intermediate
–	Vane (formerly Perplexica) ItzCrazyKns (open-source project)	Apps	Local	No public result yet	Free	advanced
–	GPT Researcher GPT Researcher (open-source project)	Search APIs & engines	Local	No public result yet	Free	advanced
–	Local Deep Research LearningCircuit (open-source project)	Apps	Local	No public result yet	Free	advanced
–	Morphic Morphic (miurla)	Apps	Local	No public result yet	Free	intermediate
–	Brave Search Brave	Apps	Cloud	No public result yet	—	beginner
–	Claude Sonnet 5.5 Anthropic	Models	Cloud	No public result yet	$4/M tokens	—
–	Consensus Consensus	Apps	Cloud	No public result yet	—	beginner
–	DeepSeek DeepSeek	Apps	Cloud	No public result yet	Free	beginner
–	Doubao ByteDance	Apps	Cloud	No public result yet	Free tier	beginner
–	Elicit Elicit	Apps	Cloud	No public result yet	Free tier · $49/mo	beginner
–	Felo Felo	Apps	Cloud	No public result yet	Free tier · $15/mo	beginner
–	Gemini 3.8 Flash Google	Models	Cloud	No public result yet	$1.5/M tokens	—
–	Google AI Mode Google	Apps	Cloud	No public result yet	Free	beginner
–	Grok 4.7 xAI	Models	Cloud	No public result yet	$3/M tokens	—
–	iAsk iAsk.Ai	Apps	Cloud	No public result yet	Free tier	beginner
–	Jina Search Foundation APIs Jina AI (Elastic)	Search APIs & engines	Cloud	No public result yet	Free tier	intermediate
–	Kimi Moonshot AI	Apps	Cloud	No public result yet	Free tier	beginner
–	Komo Komo	Apps	Cloud	No public result yet	Free tier	beginner
–	Liner Liner	Apps	Cloud	No public result yet	Free tier · $18/mo	beginner
–	Meta AI Meta	Apps	Cloud	No public result yet	Free tier	beginner
–	Metaso AI Search Metaso (秘塔科技)	Apps	Cloud	No public result yet	—	beginner
–	Qwen Chat Alibaba (Qwen team)	Apps	Cloud	No public result yet	Free	beginner
–	Sonar Perplexity AI	Models	Cloud	No public result yet	$1/M tokens	—
–	Sonar Deep Research Perplexity AI	Models	Cloud	No public result yet	$3.5/M tokens	—
–	Undermind Undermind	Apps	Cloud	No public result yet	Free tier · $16/mo	beginner
–	Valyu Valyu	Search APIs & engines	Cloud	No public result yet	—	intermediate
–	Vibe (formerly Le Chat) Mistral AI	Apps	Cloud	No public result yet	Free tier · $15/mo	beginner

Benchmarks used

• LMArena Search Arena (style control) by LMArena (arena.ai) (blind human votes, Arena score (Bradley-Terry rating from pairwise human votes on search-grounded answers, style-controlled view)) · snapshot 13 Oct 2026
  Read from LMArena's official Hugging Face dataset lmarena-ai/leaderboard-dataset (search_style_control, latest split, category overall) because arena.ai blocks automated fetches. 42 entries, all vendor search/grounding APIs (OpenAI, Anthropic, Google, xAI, Alibaba, Baidu, Perplexity, Diffbot). Current versions listed: GPT-6.1 Sol, GPT-6 Astra, GPT-6 Luna, Claude Fable 5.1, Claude Opus 5.5, Gemini 3.1 Pro, Qwen3.8-Max, Perplexity Sonar Pro / Reasoning Pro. Not listed in current versions: Claude Sonnet 5.5, Gemini 3.8 Flash, Gemini 4 Argon, Grok 4.7; no open-weights frontier models (only diffbot-small-xl). The #1 entry is the older GPT-5.6 Sol. The style-controlled view is assumed to be the site default, as for the Text Arena.
• DeepResearch Bench (RACE, GPT-5.5 judge) by DeepResearch Bench team (USTC); leaderboard hosted by muset-ai (automated tests, RACE overall score (0-100; report quality judged by GPT-5.5 relative to reference reports on 100 research tasks), ranks tools) · snapshot 11 Oct 2026
  Official DeepResearch Bench leaderboard tab 'DRB Leaderboard' (RACE judge GPT-5.5, fact-check GPT-5.4-mini), the only one maintained since 1 June 2026; the older Gemini-2.5-Pro-judge tab (45 entries, incl. Tavily, Claude Research, Kimi-Researcher, Sonar) is frozen and not comparable. Read from data_gpt55/leaderboard.csv in the Hugging Face space (last commit 2026-10-11). 17 entries: mostly research-agent submissions; the consumer products (Gemini, OpenAI, Perplexity, Grok deep research) are the original 2025 DRB paper runs re-scored with the new judge, so they reflect the 2025 products. 100 PhD-level research tasks.
• Artificial Analysis Search Index (Search APIs) by Artificial Analysis (automated tests, Equal-weighted mean of DeepSearchQA F1, BrowseComp accuracy (hard 200-task subset) and AA-Omniscience accuracy (600 private questions), 0-100, with the same answer model (GPT-5.6 Luna, medium) and harness; only the Search API varies, ranks tools) · snapshot 19 Oct 2026
  Page states 'Latest data: Oct 19, 2026'; 29 Search API configurations from 14 providers plus a model-only baseline (33.37) and OpenAI Web Search (first-party tool, 74.4). Values read from the page's embedded data (the public table shows the top 12 rows, rounded). Best tracked rows: Exa 80.31, Perplexity Search 79.98, Parallel 74.75, Brave 74.6, You.com 74.07, Firecrawl 73.49, Linkup 68.25, Tavily 65.57. Untracked: Octen (77.06), Nimble, TinyFish, DeepSeek Search, Keenable. Measured dates per row range 2026-08-17 to 2026-10-19.