Meta: Models Intelligence, Performance & Price
This analysis is intended to support you in choosing the best model provided by Meta for your use-case.
Most Intelligent
Muse Wonder (max)Muse Wonder (high)Muse Spark 1.3 (max)Muse Ember (max)Muse Ember (xhigh)Intelligence Index
Total 8 models
Lowest Time per Task
Muse Ember (xhigh)Muse Ember (max)Muse Wonder (high)Muse Spark 1.3 (max)Muse Wonder (max)Time per Intelligence Index task (minutes)
Total 8 models
Lowest Price
Muse Ember (max)Muse Ember (xhigh)Muse Spark 1.3 (max)Muse Spark 1.2 (xhigh)Muse Spark 1.3 (xhigh)Blended price (per 1M tokens)
Total 8 models
Meta offers 8 models, each with different intelligence, performance, and pricing characteristics. Below is a comparison of the key metrics across models.
- For intelligence, the top models on Meta are Muse Wonder (max) (44), Muse Wonder (high) (41), and Muse Spark 1.3 (max) (37).
- For time per task, Muse Ember (xhigh) (2.7 min), Muse Ember (max) (3.2 min), and Muse Wonder (high) (10.3 min) finish an Intelligence Index task fastest on Meta. Time per task varies significantly across models, with the slowest taking 5.2x longer than the quickest.
- For latency, Muse Ember (xhigh) (3.96s), Muse Ember (max) (5.03s), and Muse Spark 1.2 (xhigh) (8.18s) offer the lowest time to first answer token.
- For pricing, Muse Ember (max) ($0.07), Muse Ember (xhigh) ($0.07), and Muse Spark 1.3 (max) ($0.78) offer the lowest blended prices per 1M tokens. Prices vary up to 10.7x across models.
- For context window size, Muse Wonder (max) (1M), Muse Wonder (high) (1M), and Muse Spark 1.3 (max) (1M) support the largest context windows on Meta.
Intelligence
Artificial Analysis Intelligence Index
Evaluations
Agentic knowledge work, Elo
Knowledge
1 - hallucination rate
Agentic SaaS workflows
Agentic scientific research workflows in a terminal
Agentic software engineering in open-source repositories
Reasoning & knowledge
Professional document reasoning, All-pass
Agentic coding & terminal use
Physics reasoning
Quantitative analysis on spreadsheets & documents
Legal agentic work, Hallucination-Gated All-Pass Rate
Agentic business operations
Agentic real-world work tasks, Elo
Agentic interactive 3D experiences, Elo
Product and brand videos designed in code, Elo
Kubernetes incident root-cause analysis
Medical long context reasoning
Visual reasoning
Rich, varied, well-crafted language
Well-structured, scannable, answer-first
Truthful, acknowledges uncertainty, non-sycophantic
Use of filler and generic language
Intelligence Index vs. Price
Pricing
Cost per Intelligence Index Task
Intelligence Index vs. Price
Speed
Average of the single-request and multi-turn agentic workloads measured
Time per Intelligence Index Task
Quality
Context Window
Latency
Average of the single-request and multi-turn agentic workloads measured
Time to First Answer Token
Reliability
Share of requests the endpoint failed, across all the single-request and multi-turn agentic workloads measured
Endpoint Error Rate
Caching
Cache Hit Rate
Cost per Task vs. Cache Hit Rate
Further Analysis | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Muse Wonder (max) | 44 | $6.95 | 14.1 | 25.23 | 0.0% | 1M | Proprietary | |||
Muse Wonder (high) | 41 | $4.56 | 10.3 | 9.80 | 2.7% | 1M | Proprietary | |||
Muse Spark 1.3 (max) | 37 | $2.39 | 11.0 | 27.36 | 2.7% | 1M | Proprietary | |||
Muse Ember (max) | 36 | $0.14 | 3.2 | 5.03 | 0.0% | 1M | Proprietary | |||
Muse Ember (xhigh) | 34 | $0.12 | 2.7 | 3.96 | 0.0% | 1M | Proprietary | |||
Muse Spark 1.2 (xhigh) | -- | -- | -- | 8.18 | 0.0% | 1M | Proprietary | |||
Muse Spark 1.3 (xhigh) | -- | -- | -- | 13.83 | 0.0% | 1M | Proprietary | |||
Muse Spark 1.1 (xhigh) | -- | -- | -- | 13.79 | 0.0% | 1M | Proprietary | |||
Key definitions
Frequently Asked Questions
Common questions about Meta
Meta offers 8 models that we track: Muse Wonder (max), Muse Wonder (high), Muse Spark 1.3 (max), Muse Ember (max), Muse Ember (xhigh), Muse Spark 1.2 (xhigh), Muse Spark 1.3 (xhigh), and Muse Spark 1.1 (xhigh).
The most intelligent model available on Meta is Muse Wonder (max) with an Intelligence Index score of 44.
The model with the lowest time per task on Meta is Muse Ember (xhigh) at 2.7 minutes per Intelligence Index task. Time per task is the output tokens per task divided by end-to-end output speed, so lower means a task finishes sooner.
The model with the lowest time to first answer token on Meta is Muse Ember (xhigh) at 3.96s. Lower latency means faster initial response time.
The most affordable model on Meta by blended price is Muse Ember (max) at $0.0728 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Meta vary up to 40.4x across models, from $0.0728 per 1M tokens for Muse Ember (max) to $2.94 per 1M tokens for Muse Wonder (high).
Yes, Meta offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
Yes, all 8 models on Meta support JSON mode for structured output.
Yes, all 8 models on Meta support function calling (tool use).
Yes, Meta offers 8 reasoning models: Muse Wonder (max), Muse Wonder (high), Muse Spark 1.3 (max), Muse Ember (max), Muse Ember (xhigh), Muse Spark 1.2 (xhigh), Muse Spark 1.3 (xhigh), and Muse Spark 1.1 (xhigh). Reasoning models use extended thinking to work through complex problems before providing an answer.
Yes, provider performance can vary over time due to infrastructure changes, load balancing, and updates. We continuously benchmark all providers and display historical performance trends in the "Over Time" charts.
When choosing a model on Meta, consider: intelligence (for quality-sensitive tasks), time per task (how quickly the model completes an Intelligence Index task), latency (for interactive applications requiring quick first responses), pricing (for cost-sensitive workloads), and features like context window size, JSON mode, or function calling support.