Language
Independent benchmarks, pricing, latency, and provider availability for every tracked language model.
- Language models
- Model providers
The model intelligence API
Pull independent benchmarks, pricing, latency, throughput, availability, and model metadata from one documented API.
GET/api/v2/language/models/gpt-5-5-high
200OpenAI
GPT-5.5 (high)
Intelligence Index
/ 100
58.9
58.9
Coding Index
58.5
Output Speed
53.3 t/s
Input · 1M
$5.00
First Token
18.78s
{
"slug": "gpt-5-5-high",
"name": "GPT-5.5 (high)",
"model_creator": {
"name": "OpenAI"
},
"evaluations": {
"artificial_analysis_intelligence_index": 58.9,
"artificial_analysis_coding_index": 58.5
},
"pricing": {
"price_1m_input_tokens": 5,
"price_1m_output_tokens": 30
},
"performance": {
"median_output_tokens_per_second": 53.31,
"median_time_to_first_token_seconds": 18.78
}
}
Use the data behind every chart
Stop copying values from screenshots. Fetch the same benchmark, pricing, and performance data we publish on Artificial Analysis.
Output token usage per modelGET /api/v2/language/models
Intelligence Index leaderboardGET /api/v2/language/models
Intelligence vs. cost scatterGET /api/v2/language/models
What's included
Get published models, providers, benchmarks, prices, and performance metrics in a stable, documented format.
Model and provider identity
Names, creators, release dates, modalities, parameter counts, context windows, and license status for language and media models.
Composite indices
Intelligence, Coding, Agentic, Math, Openness, and Multilingual Index scores for every tracked model.
Benchmarks
Independent scores for GPQA, HLE, SciCode, IFBench, MMLU-Pro, AIME, LiveCodeBench, τ²-Bench, Terminal-Bench, AA-Omniscience, GDPval-AA, and more.
Pricing
Input, output, blended, and cache pricing per million tokens, by model and provider.
Performance distributions
Median and percentile output speed, time-to-first-token, time-to-first-answer-token, and end-to-end response time.
Time series
7-, 30-, and 90-day performance windows at daily, weekly, or monthly granularity, by model and provider, across medium, long, 100K, vision, coding, and parallel workloads.
Per-provider data
Pricing, performance, and context windows for each provider hosting a model, ready for routing and comparisons.
Provider measurements
Measurement-level exports for approved inference providers, limited to whitelisted provider slugs.
Who uses it
Teams use the API when model data needs to be current, comparable, and machine-readable.
Product teams building model selectors
Help customers choose the right model with names, creators, benchmark scores, prices, and provider availability.
Agents and model routers
Route requests to the right model and provider at runtime using current benchmark scores, prices, and availability.
Inference providers and AI labs
See how your models and endpoints perform in public benchmarks, compare them with peers, and access approved first-party measurements.
Analysts and internal AI teams
Use independent rankings, benchmark scores, and pricing across research, reports, market analysis, and internal evaluation dashboards that stay current without manual data entry.
Media generation products
Compare image, video, speech, and music systems with structured model, provider, ranking, pricing, and benchmark data.
Commercial integrations
License higher limits and redistribution rights for customer-facing products, publications, and data feeds.
One API across modalities
Integrate once, then add language, image, video, speech, and music endpoints without rebuilding your data model.
Image
Arena rankings, pricing, and capability metadata across text-to-image and image-editing systems.
- Text to image
- Image editing
Video
Pricing, runtime, and arena rankings for text-to-video and image-to-video generation models.
- Text to video
- Image to video
Audio
Quality, latency, and pricing across speech synthesis, transcription, real-time conversation, and music generation.
- Text to speech
- Speech to text
- Speech to speech
- Music generation
Built for production
Each endpoint uses stable identifiers, documented fields, pagination, rate-limit headers, and predictable errors.
Consistent shape
Structured JSON across model, provider, language, media, and evaluation endpoints.
Key-based access
Organization-scoped keys with package-aware access and documented rate limits.
Comparable metrics
Quality, speed, latency, and price fields normalized for direct comparison.
Updated continuously
Models, providers, benchmarks, and prices update as we publish them.
Production workflows
Use the same source of truth in dashboards, routing logic, analysis, and customer-facing tools.
Redistribution rights
Commercial licenses can cover redistribution in customer-facing products, reports, and data feeds.
GET/api/v2/language/models
v2# List models with benchmarks, pricing, and performance
curl "https://artificialanalysis.ai/api/v2/language/models" \
-H "x-api-key: $AA_API_KEY"Response200 OK
{"intelligence_index_version": 4,"data": [{"slug": "gpt-5-5-high","name": "GPT-5.5 (high)","model_creator": {"name": "OpenAI"},"evaluations": {"artificial_analysis_intelligence_index": 58.9,"artificial_analysis_coding_index": 58.5},"pricing": {"price_1m_input_tokens": 5,"price_1m_output_tokens": 30},"performance": {"median_output_tokens_per_second": 53.31,"median_time_to_first_token_seconds": 18.78}}]}
Choose the right access path
Use the Free API for exploration and internal workflows. Use a Commercial package for B2B integrations, higher limits, redistribution rights, package-specific media data, and approved measurement exports.
| Capability | Free API | Commercial API |
|---|---|---|
| Daily rate limit | 100 requests | Custom (1,000/day base package) |
| Model identity and headline indices | ||
| Arena rankings (image, video, speech, music) | ||
| Input and output token pricing | ||
| Median performance metrics | ||
| Individual evaluation scores | ||
| Blended pricing fields | ||
| Performance percentiles (P5 to P95) | ||
| Performance over time (7, 30, and 90 day windows) | ||
| Openness Index (8 component breakdown) | ||
| Per-provider data per model | ||
| Raw provider measurements | Approved providers only | |
| Redistribution rights | Internal use only with attribution | Commercial redistribution with attribution |
| Pricing | $0 | Package-based quote |
Self-serve
Get a free API key
Sign in, create a key, and call the public LLM and media endpoints. Internal use only; no redistribution.
Create a free API keyCommercial
Talk to us about a package
Use a Commercial package for redistribution, higher limits, package-specific media data, and approved provider or lab measurements.
Talk to salesBring Artificial Analysis data into your product
Tell us what you are building, what data you need, and whether you need internal or commercial rights. We will recommend the right package, limits, rights, and endpoints.