Anthropic has 10 models in The AI Leaderboard directory: Claude Fable 5.1, Claude Fable 5, Claude Opus 5, Claude Mythos 5.1, Claude Opus 4.8, Claude Opus 4.7, Claude Sonnet 5, Claude Mythos 5, Claude Sonnet 4.6, Claude Haiku 4.5. This provider hub compares their verified availability, licensing, context windows, pricing references and intended uses, with direct links to every model guide.
Quick answer: Claude Fable 5.1 is Anthropic's highest-ranked model here at #1 with an Overall score of 83.4. Use the lineup table below to compare it with the provider’s other current, specialist, open-weight or migration-reference models.
Anthropic model lineup at a glance
Metric | Current directory value |
|---|---|
Models covered | 10 |
Currently ranked | 3 |
Open-weight models | 0 |
Restricted models | 2 |
Legacy or migration references | 1 |
Highest-ranked model | Claude Fable 5.1, #1, Overall 83.4 |
Last verified | September 2, 2026 |
Complete Anthropic model directory
Model names in this table link directly to their complete guides. Status and pricing can change, so confirm the active endpoint and provider terms before procurement or migration.
Model | Status | License | Context | Best fit |
|---|---|---|---|---|
Available | Proprietary | 1M tokens | Long-running agents, coding, research and complex professional work | |
Available | Proprietary | 1M tokens | Long-running coding, research and complex knowledge work | |
Available | Proprietary | 1M tokens | Coding, computer use and analytical work | |
Available | Proprietary | 200,000 tokens | high-volume classification, coding subagents, chat and low-latency workflows | |
Restricted access | Proprietary | 1 million tokens | advanced cybersecurity and life-sciences research for approved organizations | |
Restricted access | Proprietary | 1 million tokens | advanced defensive cybersecurity, protein design, computational biology and specialist scientific work | |
Legacy but still documented | Proprietary | 1 million tokens | reproducing older workflows, migration testing and historical Opus comparisons | |
Available | Proprietary | 1 million tokens | coding agents, professional analysis, computer use and long-running enterprise workflows | |
Available | Proprietary | 1 million tokens | coding, computer use, agent planning and professional knowledge work | |
Available | Proprietary | 1 million tokens | high-volume agents, coding, browser work and cost-conscious professional automation |
How the Anthropic models differ
Model | Directory role | Modalities | Maximum output | Access |
|---|---|---|---|---|
Claude Fable 5.1 | Available proprietary | Text and image input; text output | 128K tokens | See the guide for current access routes |
Claude Fable 5 | Available proprietary | Text and image input; text output | 128K tokens | See the guide for current access routes |
Claude Opus 5 | Available proprietary | Text and image input; text output | 128K tokens | See the guide for current access routes |
Claude Haiku 4.5 | Available proprietary | Text and image input; text output | 64,000 tokens | Claude Platform, supported Claude products and cloud partners |
Claude Mythos 5 | Restricted | Text and image input; text output | 128,000 tokens | Invitation-only Project Glasswing and vetted Anthropic access programs |
Claude Mythos 5.1 | Restricted | Text and image input; text output | 128,000 tokens | Anthropic trusted-access programs for approved cybersecurity and life-sciences organizations |
Claude Opus 4.7 | Legacy or migration | Text and image input; text output | 128,000 tokens | Claude products, Claude API and supported cloud platforms where retained |
Claude Opus 4.8 | Available proprietary | Text and image input; text output | 128,000 tokens | Claude products, Claude API and supported cloud platforms |
Claude Sonnet 4.6 | Available proprietary | Text and image input; text output | 64,000 tokens | Claude products, Claude API, Amazon Bedrock, Google Vertex AI and Microsoft Foundry |
Claude Sonnet 5 | Available proprietary | Text and image input; text output | 128,000 tokens | Claude plans, Claude Code, Cowork, Claude API and supported cloud platforms |
The labels above describe access and lifecycle, not a hidden capability ranking. Only 3 Anthropic models in the current top eight receive a rank on this site. Other guides remain unranked until comparable evidence is available.
Anthropic model pricing
Model | Pricing reference | Cost caveat |
|---|---|---|
Claude Fable 5.1 | $10 input / $50 output per million tokens; $0.25 cache reads | Caching, tools, long context and regional rates may differ |
Claude Fable 5 | $10 input / $50 output per million tokens | Caching, tools, long context and regional rates may differ |
Claude Opus 5 | $5 input / $25 output per million tokens | Caching, tools, long context and regional rates may differ |
Claude Haiku 4.5 | $1 input and $5 output per million tokens | Caching, tools, long context and regional rates may differ |
Claude Mythos 5 | Shares Fable 5 pricing: $10 input and $50 output per million tokens | Caching, tools, long context and regional rates may differ |
Claude Mythos 5.1 | $10 input and $50 output per million tokens; access is limited to vetted organizations | Caching, tools, long context and regional rates may differ |
Claude Opus 4.7 | $5 input and $25 output per million tokens at launch | Caching, tools, long context and regional rates may differ |
Claude Opus 4.8 | $5 input and $25 output per million tokens; fast mode uses separate pricing | Caching, tools, long context and regional rates may differ |
Claude Sonnet 4.6 | $3 input and $15 output per million tokens | Caching, tools, long context and regional rates may differ |
Claude Sonnet 5 | $2 input and $10 output per million tokens; permanent pricing as of August 10, 2026 | Caching, tools, long context and regional rates may differ |
Token price is not cost per accepted result. Measure retries, tool calls, latency, infrastructure and reviewer time on the same task set.
Which Anthropic model should you choose?
Requirement | Starting point | Why |
|---|---|---|
Highest current rank | Claude Fable 5.1 | Highest Anthropic placement in the current top eight |
Open weights or self-hosting | No open-weight model in this provider directory | Compare open-weight models in the broader directory |
Restricted specialist work | Claude Mythos 5.1, Claude Mythos 5 | Designed for vetted programs and specialist safeguards |
Migration or reproduction | Claude Opus 4.7 | Retained for historical and compatibility testing |
Lowest operational cost | Measure two or three eligible models | Headline token prices omit retries, tools, caching and human review |
Anthropic models, explained
Claude Fable 5.1
Claude Fable 5.1 was released September 1, 2026. It is listed as available with proprietary. The documented context window is 1M tokens, and the maximum output is 128K tokens.
- Best fit: Long-running agents, coding, research and complex professional work.
- Modalities: Text and image input; text output.
- Access: See the guide for current access routes.
- Pricing reference: $10 input / $50 output per million tokens; $0.25 cache reads.
- Current ranking: #1 with an Overall score of 83.4.
Read the complete Claude Fable 5.1 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Fable 5
Claude Fable 5 was released June 9, 2026. It is listed as available with proprietary. The documented context window is 1M tokens, and the maximum output is 128K tokens.
- Best fit: Long-running coding, research and complex knowledge work.
- Modalities: Text and image input; text output.
- Access: See the guide for current access routes.
- Pricing reference: $10 input / $50 output per million tokens.
- Current ranking: #2 with an Overall score of 83.0.
Read the complete Claude Fable 5 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Opus 5
Claude Opus 5 was released July 24, 2026. It is listed as available with proprietary. The documented context window is 1M tokens, and the maximum output is 128K tokens.
- Best fit: Coding, computer use and analytical work.
- Modalities: Text and image input; text output.
- Access: See the guide for current access routes.
- Pricing reference: $5 input / $25 output per million tokens.
- Current ranking: #5 with an Overall score of 80.1.
Read the complete Claude Opus 5 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Haiku 4.5
Claude Haiku 4.5 was released October 15, 2025. It is listed as available with proprietary. The documented context window is 200,000 tokens, and the maximum output is 64,000 tokens.
- Best fit: high-volume classification, coding subagents, chat and low-latency workflows.
- Modalities: Text and image input; text output.
- Access: Claude Platform, supported Claude products and cloud partners.
- Pricing reference: $1 input and $5 output per million tokens.
Read the complete Claude Haiku 4.5 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Mythos 5
Claude Mythos 5 was released June 9, 2026. It is listed as restricted access with proprietary. The documented context window is 1 million tokens, and the maximum output is 128,000 tokens.
- Best fit: advanced cybersecurity and life-sciences research for approved organizations.
- Modalities: Text and image input; text output.
- Access: Invitation-only Project Glasswing and vetted Anthropic access programs.
- Pricing reference: Shares Fable 5 pricing: $10 input and $50 output per million tokens.
Read the complete Claude Mythos 5 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Mythos 5.1
Claude Mythos 5.1 was released September 1, 2026. It is listed as restricted access with proprietary. The documented context window is 1 million tokens, and the maximum output is 128,000 tokens.
- Best fit: advanced defensive cybersecurity, protein design, computational biology and specialist scientific work.
- Modalities: Text and image input; text output.
- Access: Anthropic trusted-access programs for approved cybersecurity and life-sciences organizations.
- Pricing reference: $10 input and $50 output per million tokens; access is limited to vetted organizations.
Read the complete Claude Mythos 5.1 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Opus 4.7
Claude Opus 4.7 was released April 16, 2026. It is listed as legacy but still documented with proprietary. The documented context window is 1 million tokens, and the maximum output is 128,000 tokens.
- Best fit: reproducing older workflows, migration testing and historical Opus comparisons.
- Modalities: Text and image input; text output.
- Access: Claude products, Claude API and supported cloud platforms where retained.
- Pricing reference: $5 input and $25 output per million tokens at launch.
Read the complete Claude Opus 4.7 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Opus 4.8
Claude Opus 4.8 was released May 28, 2026. It is listed as available with proprietary. The documented context window is 1 million tokens, and the maximum output is 128,000 tokens.
- Best fit: coding agents, professional analysis, computer use and long-running enterprise workflows.
- Modalities: Text and image input; text output.
- Access: Claude products, Claude API and supported cloud platforms.
- Pricing reference: $5 input and $25 output per million tokens; fast mode uses separate pricing.
Read the complete Claude Opus 4.8 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Sonnet 4.6
Claude Sonnet 4.6 was released February 17, 2026. It is listed as available with proprietary. The documented context window is 1 million tokens, and the maximum output is 64,000 tokens.
- Best fit: coding, computer use, agent planning and professional knowledge work.
- Modalities: Text and image input; text output.
- Access: Claude products, Claude API, Amazon Bedrock, Google Vertex AI and Microsoft Foundry.
- Pricing reference: $3 input and $15 output per million tokens.
Read the complete Claude Sonnet 4.6 guide for benchmarks, limitations, official sources and an evaluation framework.
Claude Sonnet 5
Claude Sonnet 5 was released June 30, 2026. It is listed as available with proprietary. The documented context window is 1 million tokens, and the maximum output is 128,000 tokens.
- Best fit: high-volume agents, coding, browser work and cost-conscious professional automation.
- Modalities: Text and image input; text output.
- Access: Claude plans, Claude Code, Cowork, Claude API and supported cloud platforms.
- Pricing reference: $2 input and $10 output per million tokens; permanent pricing as of August 10, 2026.
Read the complete Claude Sonnet 5 guide for benchmarks, limitations, official sources and an evaluation framework.
How to evaluate Anthropic models
Shortlist the Anthropic models that meet your access, licensing and modality requirements. Freeze 20 to 50 representative tasks, then give each candidate the same prompts, tools, context, reasoning budget and retry limits. Record the exact model identifier because provider aliases can change.
Measure | Record | Decision use |
|---|---|---|
Task completion | Pass rate and rubric score | Establish whether the model meets the acceptance threshold |
Reliability | Repeated-run variance and recovery behavior | Avoid choosing from one lucky result |
Quality | Factuality, instruction following and reviewer acceptance | Separate polished output from correct output |
Tool use | Wrong calls, retries, permission errors and unsafe actions | Evaluate the model and agent harness together |
Efficiency | Wall time, tokens, caching, tools and review labor | Calculate cost per accepted result |
Governance | Retention, training, regional processing and audit logs | Confirm the deployment fits policy requirements |
Availability, versions and lifecycle
Anthropic may expose different snapshots, effort settings, products and partner integrations under similar labels. Treat every model name as incomplete without an exact endpoint or checkpoint. Re-run regression tests after any model, system prompt, retrieval, tool or permission change.
- Available means a documented access route exists, not that every region or account has access.
- Preview models can change behavior, price or retirement dates before general availability.
- Open weights require an exact license and checkpoint review; “open” does not guarantee open training data or low operating cost.
- Restricted models require separate eligibility, safeguards and data-handling review.
- Legacy models remain useful for migration and reproduction but should not be the default for a new deployment.
Sources and update policy
Official Anthropic model documentation is the provider-level starting point. Each linked model guide cites its own first-party launch page, model card, pricing reference and safety documentation.
Compare this provider with the complete Best AI Models directory.
Last verified September 2, 2026. The hub should be updated when the provider launches, retires, renames or reprices a model, changes context or output limits, alters licensing, or modifies regional availability.
Frequently asked questions
How many Anthropic models are covered?
This hub covers 10: Claude Fable 5.1, Claude Fable 5, Claude Opus 5, Claude Mythos 5.1, Claude Opus 4.8, Claude Opus 4.7, Claude Sonnet 5, Claude Mythos 5, Claude Sonnet 4.6, Claude Haiku 4.5. Every model has a dedicated guide linked from the directory table and profile sections.
What is the best Anthropic model?
Claude Fable 5.1 is Anthropic's highest-ranked model in the current AI Leaderboard, at number 1 with an Overall score of 83.4. Workload-specific testing can still favor another model.
Does Anthropic offer open weights?
No open-weight Anthropic model is included in the current directory. The listed models use provider-controlled access routes.
How current is the pricing?
Pricing references were verified September 2, 2026 from first-party documentation summarized in each model guide. Always recheck the provider before budgeting because cache, batch, long-context, tool and regional rates can change.
Should every Anthropic model be tested?
No. Remove models that fail access, license, modality, context, latency or governance requirements first. Test the smallest credible shortlist on representative work.