Claude models explained: Opus, Sonnet, Haiku, and Fable Guide

Version numbers alone do not tell you where a model sits in the Claude model family. Anthropic introduced Opus 5 after Fable 5, yet still describes Fable as its most capable widely released model. Opus 5 costs at least 60% more, keeping in mind 20% more tokens consumed per task, while Haiku remains on version 4.5 and is still the fastest option in the family.

For teams putting Claude into production, those differences translate into practical choices about capability, latency, cost, and deployment.

Model lifecycle matters too: Anthropic deprecates individual model versions on published schedules, while availability can differ across deployment platforms. Amazon Bedrock and Google Cloud also set their own retirement timelines. Choosing a model is therefore not simply a matter of taking the latest release, but of matching its strengths and operating constraints to the job.

This guide traces how Claude’s tiers emerged, explains the differences that matter when choosing and deploying one, compares Claude with GPT, Gemini, and leading open-weight alternatives, and looks at where Anthropic may take the family next.

What is Claude AI?

Claude is Anthropic’s family of large language models and the name of its AI assistant and developer platform. First released publicly in March 2023, Claude is used for language, reasoning, coding, analysis, and agentic tasks, with different models offering different balances of capability, speed, and cost.

The models belong to the broader class of foundation models, general-purpose systems trained to support many downstream tasks. Claude is available through claude.ai and the Claude API, as well as on Amazon Bedrock, Google Cloud, and Microsoft Foundry.

Anthropic introduced Haiku, Sonnet, and Opus in 2024 in ascending order of capability, with speed and cost as the other major trade-offs. Fable later extended the lineup above Opus as the generally available model in the newer Mythos class.

A distinctive part of Anthropic’s approach is Constitutional AI. The company publishes the constitution used in Claude’s training; its current version was released in January 2026. Anthropic says its principles help generate synthetic conversations, responses, and preference rankings.

The approach grew out of Anthropic’s 2022 Constitutional AI research. In its original formulation, a model critiques and revises responses according to written principles, then generates AI preference data for reinforcement learning. Compared with conventional RLHF, this replaces some human preference labeling with AI feedback while keeping the principles human-written.

How the Claude model family evolved

Claude did not begin with Haiku, Sonnet, and Opus. At its public launch in March 2023, Anthropic offered Claude and the faster, lower-cost Claude Instant through its API. Later that year, Claude 2 expanded the model’s output and reasoning capabilities, while Claude 2.1 increased the context window to 200,000 tokens and added features including system prompts and early tool use. The timeline below shows the full release sequence; the rest of this section focuses on the changes that reshaped the family.

The first major structural change came with Claude 3 in March 2024. Anthropic introduced Haiku, Sonnet, and Opus together and explicitly ordered them by increasing capability, with speed and cost as the other major trade-offs. The literary names also became persistent tier labels across generations: Sonnet, for example, remained Sonnet as the version number moved on.

Three months later, Claude 3.5 Sonnet outperformed the older Claude 3 Opus across many of Anthropic’s published evaluations while retaining the speed and price of the Sonnet tier. The 3.5 generation also brought major gains in coding and vision, with Claude 3.5 Haiku following later in 2024. Anthropic subsequently described Sonnet as the preferred model for developers from June 2024 onward.

The next shift was toward longer, more autonomous work. Claude 3.7 Sonnet introduced extended thinking in February 2025 alongside the research preview of Claude Code. Claude 4 followed in May with Sonnet 4 and Opus 4, putting coding, tool use, and long-running agent workflows at the center of the launch; Opus 4.1 arrived later that year. The 4.5 generation then brought Sonnet 4.5, Haiku 4.5, and Opus 4.5.

The cadence accelerated again in 2026: Opus 4.6 and Sonnet 4.6 arrived in February, followed by Opus 4.7 in April and Opus 4.8 in May. By then, Anthropic reported that Claude authored more than 80 percent of the code merged into its own codebase, up from the low single digits before Claude Code entered research preview. That is an internal Anthropic figure rather than a general benchmark, but it shows how far Claude’s role had moved from answering prompts toward carrying out sustained work.

The family itself expanded in April 2026 with the limited-access Mythos Preview. Fable 5 and Mythos 5 followed in June, establishing a Mythos-class tier above Opus; Sonnet 5 arrived at the end of June and Opus 5 in July. The familiar Haiku–Sonnet–Opus structure remained, but it no longer defined the top of Anthropic’s lineup. For the wider chronology behind this shift, see our history of LLMs.

Claude AI models: Haiku, Sonnet, Opus, and Fable

With the lineage established, the useful distinction among current Claude models is how they spend inference budget. Haiku 4.5 is optimized for speed, while Sonnet 5 and Opus 5 use adaptive thinking and let developers adjust effort to trade capability against latency and token use. Fable 5 keeps adaptive thinking permanently enabled.

Context length also separates the generations: Haiku 4.5 has a 200,000-token window, while Sonnet 5, Opus 5, and Fable 5 support 1 million tokens.

Because prices can change, the comparisons below use relative cost positions rather than fixed figures. Check Anthropic’s current pricing before deployment.

Claude Haiku: The fastest, lowest-cost tier

Haiku 4.5 is the operational outlier in the current family. It has the shortest context window and does not use the adaptive-thinking system found in the 5-series models. Developers can instead enable extended thinking explicitly when a task needs it. For workloads built around tight feedback loops, that gives Haiku a different role from simply being a less capable Claude.

Anthropic points to chat assistants, customer-service agents, pair programming, rapid prototyping, and multi-agent systems as examples. At launch, Haiku 4.5 delivered coding performance similar to Sonnet 4 at one-third the cost and more than twice the speed, and Anthropic showed Sonnet 4.5 acting as a planner that delegated subtasks to several Haiku instances in parallel.

That makes Haiku particularly useful as a worker model when a system needs many fast calls rather than one expensive, deeply reasoned response.

Claude Sonnet: The balanced production tier

Sonnet 5 pushes the middle tier further into work that previously required larger models. It can plan, use tools such as browsers and terminals, and execute multi-step tasks autonomously. Adaptive thinking lets it vary how much reasoning it applies from turn to turn, while the effort setting gives developers another control over the cost-performance trade-off.

The important change is therefore not merely that Sonnet has become “more capable.” Its performance envelope has widened. Anthropic’s launch evaluations showed Sonnet 5 matching Opus 4.8 on some agentic tasks at higher effort while offering substantially better cost efficiency at medium effort. In practice, the same model can cover a wider range of production workloads before a team needs to move up to Opus.

Claude Opus: The high-capability core tier

For complex agentic coding and enterprise workloads, Anthropic positions Opus 5 as the starting point, with Fable aimed at workloads where maximum capability matters most. That distinction matters because Opus 5 now sits much closer to the frontier tier than the old tier names alone suggest, while retaining lower latency and cost. Opus 5 launched at the same per-token price as Opus 4.8, but in practice it consumes noticeably more tokens per task, so real inference costs run higher than 4.8's — the trade-off for its stronger position in performance rankings. 

Opus 5 also makes test-time computation part of model selection. Developers can adjust its effort level rather than changing models whenever a task needs more or less reasoning, effectively moving along a capability, latency, and token-use curve within Opus itself.

Anthropic recommends beginning at the default high setting and increasing it for the most demanding coding and agentic work. This kind of controllable inference is part of the broader shift toward reasoning in large language models.

The difference shows up most clearly on tasks where verification matters. In one Anthropic evaluation involving a real bug in an open-source package manager, Opus 5 identified the underlying cause and an additional edge case, while a competing model patched only the visible symptom.

The case illustrates the kind of work Anthropic is positioning Opus for: problems that require diagnosis and verification, not just generating a plausible fix.

Claude Fable: The Mythos-class frontier model

Fable 5 becomes most distinctive as tasks grow longer, combining always-on adaptive thinking with a large output budget. Anthropic reports that its advantage over other Claude models increases with task length and complexity, and that Fable and Mythos can operate autonomously for longer than previous Claude models.

In Stripe’s early testing, as reported by Anthropic, Fable completed a codebase-wide migration across a 50-million-line Ruby codebase in a day. Stripe estimated that the same work would otherwise have taken a team more than two months.

Choosing Fable also has an operational consequence that raw capability tables do not show: its safeguards can change which model actually answers a request. Fable runs additional classifiers for sensitive domains, and some requests are routed to a less capable model instead. Anthropic substantially reduced false-positive biology fallbacks in an August 2026 update, but dual-use areas including virology, toxicology, and molecular design still fall back to Opus 5.

The practical difference between Fable and Mythos is the safeguard layer. They use the same underlying model, but Mythos omits some of Fable’s classifiers and is available only to approved organizations through trusted-access programs such as Project Glasswing. For most developers, Fable is therefore the generally available Mythos-class model; Mythos is a restricted version for specialized work, including defensive cybersecurity.

Access briefly became part of the model’s story too. A US export-control directive led Anthropic to suspend Fable 5 and Mythos 5 on June 12, 2026. The controls were lifted later that month, and access was restored on July 1.

Claude Opus vs Sonnet vs Haiku vs Fable

The differences among current Claude models become clearest when the same workload constraints are compared side by side. Haiku minimizes latency and cost; Sonnet trades some of that efficiency for substantially more headroom; Opus prioritizes complex agentic work; and Fable extends the ceiling further for long-running, capability-sensitive tasks.

Current specifications also split the family by reasoning controls: Haiku uses manually enabled extended thinking, Sonnet and Opus use adaptive thinking, and Fable keeps adaptive thinking on by default.

The table gets you to a shortlist. From there, Anthropic recommends evaluating those models on representative prompts, data, and edge cases, then weighing the resulting quality against latency and cost. Single-run benchmark scores also hide a dimension that matters in production: whether a model succeeds every time or only on average. Toloka Arena scores models by pass^5, the probability that all five independent trials of the same task succeed, which is exactly the consistency question single-run numbers miss. For Sonnet and Opus, the effort setting adds another lever: teams can shift that balance without switching models.

A practical strategy is therefore to start at the cheapest tier that plausibly satisfies the workload and move upward only when evaluation exposes a capability gap. For tasks where failure is much more expensive than inference, the reverse can make sense: establish a quality baseline with Opus or Fable, then test whether Sonnet can reproduce it more efficiently.

Claude vs other AI model families

The snapshot below reflects public model lineups in August 2026. Benchmark positions change quickly, so our LLM leaderboard overview is the better place for current rankings, and Toloka Arena tracks how Claude, GPT, and Gemini models perform on private, non-contaminated agentic benchmarks, scored by pass^5 reliability and cost per task. Because Arena tasks are never in any model’s training data, the standings show how these families actually behave on unseen work rather than on memorized tests. The comparison here focuses mainly on differences in how these families are packaged, deployed, and controlled.

Claude vs ChatGPT and GPT models

For teams deploying AI systems, the closer comparison is Claude versus the GPT model family rather than Claude versus ChatGPT, which is the consumer-facing app.

OpenAI’s current GPT-5.6 lineup now resembles Anthropic’s tier structure surprisingly closely: Sol is the frontier model, Terra balances capability and cost, and Luna targets high-volume, cost-sensitive workloads.

Benchmark leadership remains task-dependent. As of July 24, 2026, Anthropic reported Opus 5 ahead of all other models on Frontier-Bench v0.1, a software-engineering evaluation. OpenAI’s July 9 results showed Mythos 5 narrowly ahead of GPT-5.6 Sol on the 1-million-token GraphWalks BFS test, while Fable 5 led Sol on GDPval-AA v2, an evaluation of professional knowledge work.

Claude vs Gemini

Google’s Gemini lineup combines general-purpose models with a broader set of specialized variants for different modalities and workloads. Its current API includes Flash and Flash-Lite variants alongside Pro, plus separate models for image generation, real-time interaction, translation, and other multimodal tasks.

That breadth can matter when audio, image, or real-time interaction is central to the application rather than an added capability around a primarily text-and-code model.

Claude vs Grok

xAI’s Grok lineup is narrower and more concentrated around coding, agentic tasks, and knowledge work. Grok 4.5 targets those workloads, while Grok Build extends the lineup into autonomous software-engineering workflows. The overlap with Claude is substantial, but xAI does not divide its models as explicitly into persistent capability tiers.

Claude vs open-weight models

The bigger architectural choice appears when Claude is compared with open-weight families such as Llama, DeepSeek, Qwen, and GLM. Many models in these families make their weights available for teams to deploy themselves, modify where the license permits, and integrate into infrastructure they control.

The exact freedoms depend on the model’s license, which is why “open-weight” is often more precise than “open-source.” Llama, DeepSeek, Qwen, and GLM all have openly released model weights, but under different licensing terms.

That changes the engineering trade-off. With Claude, model serving and vendor-side safeguards remain with Anthropic; self-hosted models give teams more control over infrastructure, fine-tuning, and data flow, while shifting more responsibility for deployment and maintenance to the team running them. Our guide to open-source vs closed-source LLMs goes deeper into that distinction.

Anthropic models: What comes next for Claude?

Anthropic has not announced a named successor to Opus 5. The progression from Claude 3 to 3.5, 4, 4.5, and 5, together with the rapid 4.x releases in 2026, suggests that the company is likely to keep iterating quickly.

Opus 5 was launched specifically around long-running agents, while Anthropic’s own research shows Claude taking on an increasing share of the work involved in developing AI systems themselves.

The model is also becoming only one layer of the product. Claude Science, for example, combines Claude with research tools, compute, and auditable artifacts rather than treating the model as a standalone chat interface. Anthropic has taken similar domain-specific approaches in software development and cybersecurity.

At the frontier, access may become more differentiated as capability rises. Fable and Mythos already show that the same underlying model can be offered under different safeguard and access regimes, while Project Glasswing gives selected organizations access to capabilities Anthropic considers too sensitive for unrestricted release. Context capacity and safety controls are also likely to remain active areas of development.

How human data supports model training and evaluation

Constitutional AI automates part of the preference-labeling process, but Anthropic still relies on people to design evaluations, investigate failures, and red-team its models. Its system cards combine automated testing with human and expert review, especially for behavior that is hard to capture with a single metric.

Human input matters during training and fine-tuning when the target behavior cannot be captured by a simple rule or automated score. Domain experts can provide examples, preference judgments, and annotations for specialist tasks, while gaps in that data can later show up as uneven model performance.

Automated evaluators can make testing easier to scale, but teams still use human judgments to check what those scores actually capture. With agents, that often means evaluating intermediate steps such as plans, tool calls, and error recovery rather than looking only at the final answer.

Human data is therefore most valuable where the target is hard to measure automatically: specialist knowledge, subjective quality, unusual failure modes, and safety boundaries. Toloka provides data for LLM training and fine-tuning, expert model evaluation, and AI safety testing for those parts of the development process.

Frequently asked questions

What is the Claude model family?

Claude is Anthropic’s family of general-purpose AI models. The names Haiku, Sonnet, Opus, and Fable identify different tiers, while numbers such as 4.5 or 5 identify generations within those tiers.

Why is Claude Haiku still on version 4.5?

Anthropic does not update every tier at the same time. Haiku 4.5 remains the current Haiku model while Sonnet, Opus, and Fable have moved to generation 5. The version number therefore tells you when a model belongs in Claude’s development history, not where it ranks in the current lineup.

Which Claude model is best for coding?

There is no single coding tier for every workload. Sonnet 5 is designed for strong agentic coding with better cost efficiency, while Anthropic positions Opus 5 for more complex coding and knowledge work; Fable 5 pushes further into ambitious, long-running projects that may run for hours or days.

Which Claude model is best for agentic tool use?

It depends on how much reliability the workflow needs. Toloka Arena evaluates Claude models alongside GPT and Gemini on private, multi-turn tool-use tasks with strict business rules, scoring each model by pass^5, the probability that all five independent runs of a task succeed. Checking the current Arena standings shows how each Claude tier balances tool-use reliability against inference cost per task before you commit to one in production — and since standings differ across Arena's domains, from airline and manufacturing scenarios to banking HR and logistics, it's worth checking the domain closest to your own workload rather than the composite score alone 

Is Fable replacing Opus?

No. Fable extends the Claude lineup above Opus rather than replacing it. Anthropic describes Fable 5 as its most capable generally available model, while Opus remains a separate tier for demanding work where Fable’s additional capability and cost are unnecessary.

What is the difference between Fable and Mythos?

Fable 5 and Mythos 5 use the same underlying model. Fable adds safeguards that make Mythos-level capability suitable for general availability; Mythos removes some of those restrictions and is available only through trusted-access programs such as Project Glasswing.

From model choice to system design

The lesson from Claude’s expanding family is that model selection increasingly happens at the task level. A single project may need fast, inexpensive inference in one part of the workflow, deeper reasoning in another, and a frontier model only for a small set of difficult cases. The practical question is therefore how to build the right model architecture for the project.

Claude makes that pattern unusually visible because its tiers span different operating profiles, but the same logic applies across model families. A project can route routine work to cheaper models and reserve stronger ones for tasks where evaluation shows a measurable gain.

Those routing decisions are only as good as the evaluation data behind them. Teams need representative prompts, edge cases, and human judgments that reflect real failure modes, and ideally independent evaluation on tasks models have never trained on. Otherwise, the model that looks best in testing may not be the one that performs best in production.

Related reading

Base LLM vs instruction-tuned LLM

Transformer architecture

Fable 5 reset the leaderboard. The blind spots didn’t move.