Where does the leaderboard data come from?

Benchmark data comes from Artificial Analysis, an independent LLM benchmarking service. Quality scores, output speed, and time-to-first-token are measured by their infrastructure and refreshed every 6 hours.

What is the Quality Index?

The Quality Index is a composite score from Artificial Analysis that aggregates performance across multiple benchmarks including reasoning, coding, math, and general knowledge tasks.

What does the Value ranking show?

The Value ranking combines quality score with pricing to show which models give you the most capability per dollar. A high-value model delivers strong benchmarks at a low price per token.

How often is the leaderboard updated?

Benchmark and pricing data refreshes every 6 hours. New models are added automatically as they appear in our data sources.

Can I see latency metrics?

Yes. The Speed ranking shows output tokens per second and time-to-first-token (TTFT) for each model, helping you choose models for latency-sensitive applications.

LLM Leaderboard

Live rankings by quality, speed, and value. Data from Artificial Analysis benchmarks.

Composite intelligence score from Artificial Analysis benchmarks

GPT-5.3 Codex

OpenAI

idx

GPT-5.4

OpenAI

idx

GPT-5.2

OpenAI

idx

Quality vs Price

Higher and left = better value. Hover for details.

OpenAI

Anthropic

Google

xAI

Meta

Mistral

Amazon

NVIDIA

Cohere

Perplexity

Moonshot

Zhipu

MiniMax

#	Model	Provider	Quality Index	Speed	Input/1M	Context
1	GPT-5.4OpenAI	OpenAI	39idx	0 tok/s	$2.5	1.1M
2	GPT-5.3 CodexOpenAI	OpenAI	33idx	115 tok/s	$1.75	400K
3	GPT-5.2OpenAI	OpenAI	30idx	0 tok/s	$1.75	400K
4	Gemini 3.1 ProGoogle	Google	30idx	138 tok/s	$2	1.0M
5	GLM-5Zhipu	Zhipu	28idx	0 tok/s	$1	200K
6	Claude Opus 4.6Anthropic	Anthropic	26idx	0 tok/s	$5	1.0M
7	Grok 4.20xAI	xAI	26idx	0 tok/s	$1.25	1.0M
8	GPT-5.1OpenAI	OpenAI	25idx	0 tok/s	$1.25	400K
9	Claude Sonnet 4.6Anthropic	Anthropic	25idx	0 tok/s	$3	1.0M
10	Claude Opus 4.5Anthropic	Anthropic	24idx	0 tok/s	$5	200K
11	Kimi K2.5Moonshot	Moonshot	24idx	0 tok/s	$0.6	262K
12	GPT-5OpenAI	OpenAI	23idx	0 tok/s	$1.25	400K
13	MiniMax M2.5MiniMax	MiniMax	23idx	0 tok/s	$0.3	128K
14	o3OpenAI	OpenAI	20idx	155 tok/s	$2	200K
15	Gemini 3 FlashGoogle	Google	18idx	0 tok/s	$0.5	1.0M
16	GPT-5 MiniOpenAI	OpenAI	17idx	0 tok/s	$0.25	400K
17	o4 MiniOpenAI	OpenAI	17idx	0 tok/s	$1.1	200K
18	Gemini 2.5 ProGoogle	Google	16idx	0 tok/s	$1.25	1.0M
19	Gemini 3.1 Flash-LiteGoogle	Google	16idx	0 tok/s	$0.25	1.0M
20	Claude Haiku 4.5Anthropic	Anthropic	15idx	0 tok/s	$1	200K
21	o1OpenAI	OpenAI	15idx	0 tok/s	$15	200K
22	Nova 2.0 Pro ReasoningAmazon	Amazon	14idx	136 tok/s	$1.25	128K
23	GPT-5 NanoOpenAI	OpenAI	13idx	0 tok/s	$0.05	400K
24	Nemotron 3 Super 120BNVIDIA	NVIDIA	13idx	0 tok/s	$0.3	1.0M
25	GPT-4.1OpenAI	OpenAI	13idx	0 tok/s	$2	1.0M
26	o3 MiniOpenAI	OpenAI	13idx	0 tok/s	$1.1	200K
27	GPT-4.1 MiniOpenAI	OpenAI	10idx	0 tok/s	$0.4	1.0M
28	Gemini 2.5 FlashGoogle	Google	10idx	0 tok/s	$0.3	1.0M
29	Llama 4 MaverickMeta	Meta	9idx	44 tok/s	$0.27	1.0M
30	Mistral Large 3Mistral	Mistral	9idx	0 tok/s	$0.5	262K
31	Nova 2.0 LiteAmazon	Amazon	9idx	237 tok/s	$0.3	1.0M
32	GPT-4oOpenAI	OpenAI	8idx	0 tok/s	$2.5	128K
33	GPT-4.1 NanoOpenAI	OpenAI	8idx	0 tok/s	$0.1	1.0M
34	Sonar ProPerplexity	Perplexity	8idx	0 tok/s	$3	128K
35	Command ACohere	Cohere	7idx	58 tok/s	$2.5	128K
36	GPT-4o MiniOpenAI	OpenAI	7idx	0 tok/s	$0.15	128K
37	Gemini 2.5 Flash-LiteGoogle	Google	7idx	0 tok/s	$0.1	1.0M
38	Llama 4 ScoutMeta	Meta	7idx	131 tok/s	$0.08	1.0M

Rankings based on live benchmark data. Quality = composite intelligence index. Value = quality index / input cost per 1M tokens. Latency = time to first token.

Data by Artificial Analysis

How to Use the LLM Leaderboard

1
Choose a ranking metric
Switch between Quality, Speed, and Value tabs to rank models by the metric that matters most to your use case.
2
Filter by provider
Use the provider buttons to focus on specific vendors. Compare only OpenAI models, or pit Anthropic against Google.
3
Explore the scatter chart
The interactive quality-vs-price chart plots every model so you can visually identify the best value picks.

Why Use This Leaderboard

Three ranking modes (Quality, Speed, and Value) for different decision criteria
Benchmark data from Artificial Analysis, refreshed every 6 hours
Interactive scatter chart plotting quality against cost for visual comparison
Provider filtering to narrow the field to vendors you're evaluating
Includes output speed (tokens/sec) and time-to-first-token for latency planning

Common Use Cases

Model selection

Find the highest-quality model within your budget by sorting on the Value tab.

Latency optimization

Sort by Speed to find the fastest models for real-time applications like chat or autocomplete.

Benchmark tracking

Check back regularly to see how new model releases stack up against existing options.

Stakeholder reporting

Use the scatter chart to show leadership why a specific model offers the best quality-to-cost ratio.

Related Tools

Token Counter

Count tokens and see per-request costs.

API Pricing

Full pricing table for all models.

Compare Models

Side-by-side model comparison.

Cost Calculator

Estimate monthly API spend.

Frequently Asked Questions

Common questions about the LLM leaderboard

LLM Leaderboard

Quality vs Price

How to Use the LLM Leaderboard

Choose a ranking metric

Filter by provider

Explore the scatter chart

Why Use This Leaderboard

Common Use Cases

Model selection

Latency optimization

Benchmark tracking

Stakeholder reporting

Related Tools

Frequently Asked Questions