Anthropic: Models Intelligence, Performance & Price
Analysis of Anthropic's models across key metrics including quality, price, output speed, latency, context window & more. This analysis is intended to support you in choosing the best model provided by Anthropic for your use-case.
Most Intelligent
Intelligence index
Total 26 models
Fastest
Output speed
Total 26 models
Lowest Price
Blended price (per 1M tokens)
Total 26 models
Anthropic offers 26 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 Anthropic are Claude Opus 5 (max) (63), Claude Opus 5 (xhigh) (63), and Claude Fable 5 (with fallback) (62).
- For output speed, the fastest models are Claude 4.5 Haiku (98 t/s), Claude 4.5 Haiku (88 t/s), and Claude Sonnet 5 (xhigh) (82 t/s).
- For latency, Claude 4.5 Haiku (0.90s), Claude Opus 4.5 (1.37s), and Claude Opus 4.7 (Non-reasoning, high) (1.50s) offer the lowest time to first answer token.
- For pricing, Claude 4.5 Haiku ($0.77), Claude 4.5 Haiku ($0.77), and Claude Sonnet 5 (max) ($1.54) offer the lowest blended prices per 1M tokens.
- For context window size, Claude Opus 5 (max) (1M), Claude Opus 5 (xhigh) (1M), and Claude Fable 5 (with fallback) (1M) support the largest context windows on Anthropic.
- Claude 4.5 Haiku offers both the fastest output and best pricing, making it attractive for throughput-sensitive and cost-conscious applications. Claude Opus 5 (max) leads in intelligence for tasks that require the highest quality.
Highlights
Intelligence Evaluations
Artificial Analysis Intelligence Index
Intelligence Evaluations
Agentic real-world work tasks, (Elo-500)/2000
Agentic tool use
Agentic coding & terminal use
Coding
Reasoning & knowledge
Scientific reasoning
Physics reasoning
Knowledge
1 - hallucination rate
Long context reasoning
Agentic knowledge work, Elo
Agentic SaaS workflows
Legal agentic work, criterion pass rate
Agentic business operations
Instruction following
Long-horizon agentic tasks
Kubernetes incident root-cause analysis
Visual reasoning
Intelligence Index vs. Price
Context Window
Context Window
Pricing
Intelligence Index vs. Price
Performance Summary
Output Speed vs. Price
Speed
Measured by Output Speed (tokens per second)
Output Speed
Latency
Measured by Time (seconds) to First Token
Latency: Time To First Answer Token
End-to-End Response Time
Seconds to output 500 tokens, calculated based on time to first token, 'thinking' time for reasoning models, and output speed
End-to-End Response Time vs. Price
Further Analysis | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
Claude Opus 5 (max) | 1M | Proprietary | 63 | $2.34 | 57 | 67.27 | 76.03 | -- | |||
Claude Opus 5 (xhigh) | 1M | Proprietary | 63 | $1.80 | 58 | 24.54 | 33.15 | -- | |||
Claude Fable 5 (with fallback) | 1M | Proprietary | 62 | $3.14 | 69 | 137.13 | 144.38 | -- | |||
Claude Opus 5 (high) | 1M | Proprietary | 61 | $1.23 | 57 | 10.67 | 19.37 | -- | |||
Claude Opus 5 (medium) | 1M | Proprietary | 59 | $0.72 | 57 | 5.92 | 14.68 | -- | |||
Claude Opus 4.8 (max) | 1M | Proprietary | 57 | $2.03 | 60 | 20.33 | 28.61 | -- | |||
Claude Sonnet 5 (max) | 1M | Proprietary | 55 | $1.72 | 73 | 170.85 | 177.65 | -- | |||
Claude Opus 4.7 (max) | 1M | Proprietary | 55 | $2.23 | 52 | 14.89 | 24.51 | -- | |||
Claude Opus 5 (low) | 1M | Proprietary | 52 | $0.43 | 56 | 3.22 | 12.14 | -- | |||
Claude Sonnet 4.6 (max) | 1M | Proprietary | 48 | $1.22 | 50 | 166.72 | 176.64 | -- | |||
Claude Opus 4.6 (max) | 1M | Proprietary | 45* | -- | 43 | 12.41 | 24.00 | -- | |||
Claude Opus 4.7 (Non-reasoning, high) | 1M | Proprietary | 44* | -- | 50 | 1.50 | 11.47 | -- | |||
Claude Sonnet 5 (Non-reasoning) | 1M | Proprietary | 43 | $0.42 | 72 | 1.81 | 8.78 | -- | |||
Claude Opus 4.5 | 200k | Proprietary | 42* | -- | 49 | 10.12 | 20.23 | -- | |||
Claude Opus 4.6 (high) | 1M | Proprietary | 39* | -- | 40 | 2.22 | 14.75 | -- | |||
Claude 4.5 Sonnet | 1M | Proprietary | 37 | $0.46 | 42 | 11.75 | 23.75 | -- | |||
Claude Sonnet 4.6 (Non-reasoning) | 1M | Proprietary | 37* | -- | 43 | 1.76 | 13.39 | -- | |||
Claude Opus 4.5 | 200k | Proprietary | 36* | -- | 49 | 1.37 | 11.62 | -- | |||
Claude Sonnet 4.6 (Non-reasoning, Low Effort) | 1M | Proprietary | 35* | -- | 44 | 1.75 | 13.24 | -- | |||
Claude 4.1 Opus | 200k | Proprietary | 35* | -- | -- | -- | -- | -- | |||
Claude 4 Opus | 200k | Proprietary | 32* | -- | -- | -- | -- | -- | |||
Claude 4.5 Sonnet | 1M | Proprietary | 30* | -- | 41 | 1.60 | 13.90 | -- | |||
Claude 4.5 Haiku | 200k | Proprietary | 30 | $0.22 | 98 | 17.73 | 22.82 | -- | |||
Claude 4 Sonnet | 1M | Proprietary | 30 | -- | -- | -- | -- | -- | |||
Claude 4.1 Opus | 200k | Proprietary | 29* | -- | -- | -- | -- | -- | |||
Claude 4 Opus | 200k | Proprietary | 26* | -- | -- | -- | -- | -- | |||
Claude 4 Sonnet | 1M | Proprietary | 26* | -- | -- | -- | -- | -- | |||
Claude 4.5 Haiku | 200k | Proprietary | 24* | -- | 88 | 0.90 | 6.55 | -- | |||
Claude 3 Opus | 200k | Proprietary | 12* | -- | -- | -- | -- | -- | |||
Claude Sonnet 5 (low) | 1M | Proprietary | -- | -- | 67 | 2.60 | 10.10 | -- | |||
Claude Sonnet 5 (xhigh) | 1M | Proprietary | -- | -- | 82 | 51.53 | 57.65 | -- | |||
Claude Sonnet 5 (high) | 1M | Proprietary | -- | -- | 76 | 8.50 | 15.05 | -- | |||
Claude Sonnet 5 (medium) | 1M | Proprietary | -- | -- | 69 | 3.24 | 10.53 | -- | |||
Key definitions
Frequently Asked Questions
Common questions about Anthropic
Anthropic offers 26 models that we track: Claude Opus 5 (max), Claude Opus 5 (xhigh), Claude Fable 5 (with fallback), Claude Opus 5 (high), Claude Opus 5 (medium), Claude Opus 4.8 (max), Claude Sonnet 5 (max), Claude Opus 4.7 (max), Claude Opus 5 (low), Claude Sonnet 4.6 (max), Claude Opus 4.6 (max), Claude Opus 4.7 (Non-reasoning, high), Claude Sonnet 5 (Non-reasoning), Claude Opus 4.5, Claude Opus 4.6 (high), Claude 4.5 Sonnet, Claude Sonnet 4.6 (Non-reasoning), Claude Opus 4.5, Claude Sonnet 4.6 (Non-reasoning, Low Effort), Claude 4.5 Sonnet, Claude 4.5 Haiku, Claude 4.5 Haiku, Claude Sonnet 5 (low), Claude Sonnet 5 (xhigh), Claude Sonnet 5 (high), and Claude Sonnet 5 (medium).
The most intelligent model available on Anthropic is Claude Opus 5 (max) with an Intelligence Index score of 63.
The fastest model on Anthropic by output speed is Claude 4.5 Haiku at 98.2 tokens per second.
The model with the lowest time to first answer token on Anthropic is Claude 4.5 Haiku at 0.90s. Lower latency means faster initial response time.
The most affordable model on Anthropic by blended price is Claude 4.5 Haiku at $0.77 per 1M tokens (7:2:1 cache hit/input/output ratio).
Prices on Anthropic vary up to 10x across models, from $0.77 per 1M tokens for Claude 4.5 Haiku to $7.70 per 1M tokens for Claude Fable 5 (with fallback).
Yes, Anthropic offers an OpenAI-compatible API, making it easy to switch from OpenAI or use existing OpenAI SDK integrations.
9 of 26 models on Anthropic support JSON mode for structured output.
Yes, all 26 models on Anthropic support function calling (tool use).
Yes, Anthropic offers 18 reasoning models: Claude Opus 5 (max), Claude Opus 5 (xhigh), Claude Fable 5 (with fallback), Claude Opus 5 (high), Claude Opus 5 (medium), Claude Opus 4.8 (max), Claude Sonnet 5 (max), Claude Opus 4.7 (max), Claude Opus 5 (low), Claude Sonnet 4.6 (max), Claude Opus 4.6 (max), Claude Opus 4.5, Claude 4.5 Sonnet, Claude 4.5 Haiku, Claude Sonnet 5 (low), Claude Sonnet 5 (xhigh), Claude Sonnet 5 (high), and Claude Sonnet 5 (medium). 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 Anthropic, consider: intelligence (for quality-sensitive tasks), output speed (for throughput-intensive tasks), 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.