Anthropic

10 models from Anthropic ranked on LMRank.

Anthropic models
Rank Model Tags Context Price (in / out)
1 Claude Fable 5
Agentic Coding Intelligent Reasoning Long Context
1,000,000 $10.00 / $50.00
2 Claude Opus 4.8
Intelligent Reasoning Agentic Multimodal
1,000,000 $15.00 / $75.00
3 Claude Opus 4.7
Intelligent Reasoning Agentic Multimodal
200,000 $15.00 / $75.00
5 Claude Opus 4.5
Intelligent Reasoning Multimodal
200,000 $15.00 / $75.00
7 Claude Sonnet 5
Intelligent Reasoning Coding Agentic Image Input Long Context
1,000,000 $2.00 / $10.00
14 Claude Opus 4.6
Intelligent Coding Agentic Long Context
1,000,000 $5.00 / $25.00
17 Claude Sonnet 4.6
Agentic Coding Intelligent
1,000,000 $3.00 / $15.00
24 Claude Sonnet 4
Intelligent Coding Multimodal
200,000 $3.00 / $15.00
55 Anthropic: Claude Haiku 4.5
Fast Coding
200,000 $1.00 / $5.00
85 Claude Haiku 4
Intelligent Fast Coding
200,000 $0.25 / $1.25

In Depth: Anthropic

About Anthropic

Anthropic is an AI research lab dedicated to building safe, steerable large language models. Their flagship Claude Fable 5 is currently ranked #1 on LMRank with a 9.9/10, making it a top choice for enterprises prioritizing controlled, reliable outputs.

Anthropic releases models at a measured cadence, favoring safety evaluations over speed. All current models are closed-weight and available exclusively via Anthropic API, Claude.ai, and select enterprise contracts. The company does not offer open-weight releases, which limits customization but supports consistent, managed behavior.

Model Lineup

Anthropic offers 9 active models. Their top performer is Claude Fable 5, ranked #1 on LMRank at 9.9/10. The lineup spans reasoning-focused, cost-balanced, and high-efficiency tiers for different workload demands.

The top tier includes Claude Fable 5 at 9.9/10. Mid-tier models like Claude Opus 3 and Claude Sonnet 4 balance performance with lower latency. Claude Haiku 3 is the lightweight entry point for simpler tasks. All models are closed-weight and accessed via the same API, with tiered pricing by capability.

Pricing & Access

Anthropic models require API or web app access; no free tier exists for the highest-performance options. Prices start with pay-as-you-go rates for Claude Haiku 3 and scale up for Claude Fable 5 by input/output token.

Claude Fable 5 is priced at market-leading rates per token, similar to OpenAI’s GPT-4.5 tier. Claude Haiku 3 is the most affordable. Access is restricted to Anthropic’s proprietary API and Claude.ai; no open-weight or self-hosted option is available. Enterprise discounts and dedicated throughput are negotiable.

Best For & When to Look Elsewhere

For safety-critical applications requiring high reliability, Claude Fable 5 sets the benchmark. Claude Haiku 3 is best for high-volume, low-cost inference. Anthropic models are a poor fit if you need open weights or on-premise deployment.

  • Enterprise chatbots needing strict content guardrails
  • Complex reasoning tasks with detailed oversight
  • Use any other lab if you require open-weight models