← All providers
ANTHROPIC / MODEL DIRECTORY

Anthropic model matrix

Choose the model type. Compare its position. See when it arrived.

Compare Anthropic models in our directory by task, positioning and release date.

15models
01model types
03selection tiers

Recent model landscape

10

The 10 most recent dated models · time × selection tier

Swipe to explore the timeline. Tap a model for details.

FlagshipMainstreamValue2025-10-152026-01-122026-04-122026-07-102026-10-07Time →Claude Opus 4.7 · 2026-04-16 · Text · FlagshipClaude Opus 4.7Text · 04-16Claude Opus 4.8 · 2026-05-28 · Text · FlagshipClaude Opus 4.8Text · 05-28Claude Fable 5 · 2026-06-09 · Text · FlagshipClaude Fable 5Text · 06-09Claude Opus 5 · 2026-07-24 · Text · FlagshipClaude Opus 5Text · 07-24Claude Fable 5.1 · 2026-09-01 · Text · FlagshipClaude Fable 5.1Text · 09-01Claude Sonnet 4.6 · 2026-02-17 · Text · MainstreamClaude Sonnet 4.6Text · 02-17Claude Sonnet 5 · 2026-06-30 · Text · MainstreamClaude Sonnet 5Text · 06-30Claude Sonnet 5.5 · 2026-09-28 · Text · MainstreamClaude Sonnet 5.5Text · 09-28Claude Haiku 4.5 · 2025-10-15 · Text · ValueClaude Haiku 4.5Text · 10-15Claude Haiku 5.5 · 2026-10-07 · Text · ValueClaude Haiku 5.5Text · 10-07
Claude Haiku 5.5ValueModel announcement ↗
Model details ↗

Horizontal distance follows calendar time. Vertical bands are selection tiers, not benchmark scores; labels may occupy separate tracks within a tier. Select a model to inspect its source. 5 models without a date or tier are excluded.

Input price per 1M tokens: list vs Crazyrouter
ListCrazyrouterClaude Fable 5$10 → $6.50Claude Opus 5$5.00 → $3.25Claude Opus 4.8$5.00 → $3.25Claude Opus 4.7$5.00 → $3.25Claude Sonnet 4.6$3.00 → $1.95Claude Sonnet 5.5$2.00 → $2.00Claude Sonnet 5$2.00 → $1.30Claude Haiku 4.5$1.00 → $0.650Claude Haiku 5.5$0.100 → $0.100

Complete model matrix

Text

15 models shown

Dates follow official model or family announcements. API availability may differ; unverified records are labeled.

Scroll horizontally to compare tiers.

Read the matrix. Columns show positioning, rows show release months. Blank cells mean no listed model in that combination. Unknown dates are always placed last.
Release month
FlagshipCapability first
MainstreamBalanced everyday workloads
ValueCost & throughput first
Not classifiedPositioning under review
2026101 models
—
—
—
2026092 models
—
—
2026071 models
—
—
—
2026062 models
—
—
2026051 models
—
—
—
2026041 models
—
—
—
2026031 models
—
—
—
claude-opus-4-6
Unverified

Listed on Pricing; exact-version positioning awaits verification.

Input$3.25Output$16.25/ 1M tokens
Not classifiedPricing ↗
2026021 models
—
—
—
2026012 models
—
—
—
claude-opus-4-5
Unverified

Listed on Pricing; exact-version positioning awaits verification.

Input$3.25Output$16.25/ 1M tokens
Not classifiedPricing ↗
claude-sonnet-4-5
Unverified

Listed on Pricing; exact-version positioning awaits verification.

Input$1.95Output$9.75/ 1M tokens
Not classifiedPricing ↗
2025101 models
—
—
—
Date unverified2 models
—
—
claude-fable-5-2
Date unverified

Listed on Pricing; exact-version positioning awaits verification.

Input$6.8Output$34/ 1M tokens
Not classifiedPricing ↗

Columns show positioning, rows show release months. Blank cells mean no listed model in that combination. Unknown dates are always placed last.

Dates prefer official model or family announcements. Where only API availability is verified, it is explicitly labeled API release, not a model debut. Directory dates are unverified. Missing dates are never inferred from names or listing dates; generic endpoints and unresolved aliases are not assigned a version launch date.

Classification reviewed

How the tiers work

These are Crazyrouter selection tiers for Anthropic, not the vendor's official ratings. Compare within the same category and generation; newer or more expensive does not automatically mean a better fit.

Classification reviewed

Categories follow output task: text, image, video, embeddings and audio/music are compared separately; price is a cross-check, not a capability score.

Flagship: capability, complex tasks, quality or control come first; sharing a tier does not imply equal capability.

Mainstream: balances capability, latency and cost for production; speed and balanced variants are classified by use, not price alone.

Value: cost, latency or throughput come first; lower cost does not imply better quality on every task.

Coding, reasoning, speed and specialist capabilities are descriptive tags; output modalities are assessed separately, with video and audio prices read alongside specs and duration.

Anthropic: model positioning & official evidence

Fable and Opus are flagship, Sonnet mainstream, Haiku value. Compare standard uncached, non-Batch rates; Haiku 5.5 has a 100K input threshold. The official Opus 5.5 news index and article show conflicting dates, so neither is treated as verified.

Official reference rates below are separate from Crazyrouter prices. Price is a cross-check, not a capability score. Unverified prices are left unspecified.

  • Claude Opus 5.5Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 4, output 20. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Sonnet 5.5Text · Mainstream

    Sonnet balances coding and agent capability with everyday cost.

    USD / 1M tokens: input 2, output 10. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Haiku 5.5Text · Value

    Haiku prioritizes latency, volume and cost.

    USD / 1M tokens: input 0.1, output 0.5. Input ≤100K; above 100K: $0.50 / $2.50. Official pricing ↗

    Official source ↗
  • Claude Fable 5.1Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 10, output 50. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Fable 5Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 10, output 50. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Opus 5Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 5, output 25. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Opus 4.8Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 5, output 25. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Opus 4.7Text · Flagship

    Opus / Fable occupy the high-capability band; sharing a tier does not imply equal capability.

    USD / 1M tokens: input 5, output 25. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Sonnet 5Text · Mainstream

    Sonnet balances coding and agent capability with everyday cost.

    USD / 1M tokens: input 2, output 10. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Sonnet 4.6Text · Mainstream

    Sonnet balances coding and agent capability with everyday cost.

    USD / 1M tokens: input 3, output 15. Standard; uncached input. Official pricing ↗

    Official source ↗
  • Claude Haiku 4.5Text · Value

    Haiku prioritizes latency, volume and cost.

    USD / 1M tokens: input 1, output 5. Standard; uncached input. Official pricing ↗

    Official source ↗

Guides and comparisons

Podcast

AI Dev Tools: The Crazyrouter Podcast

Short episodes for developers on AI APIs, gateways, cost control and coding tools.

Latest episodes

  • EP196 · Oct 29, 2026 · 6:53

    AI API Security Posture — Turn Controls Into Continuous Evidence

    A practical guide to AI API security posture: inventory exposure, verify controls continuously, protect keys and data, constrain tools, and turn security assumptions into evidence that operators can act on.

  • EP195 · Oct 28, 2026 · 7:21

    AI API Policy Drift — Detect When Production Rules Stop Matching Intent

    A practical guide to policy drift in AI APIs: version routing and safety rules, detect stale enforcement, compare intended and observed behavior, and restore control before small gaps become incidents.

Episodes with written show notes →

Questions & answers

Is this the real Claude model?
Yes. Requests are routed to Anthropic-backed channels and billed per token. Every request's upstream route and token count are visible in the console log; you can verify with Anthropic's own tokenizer behaviour and the `usage` block.
How much does the Claude API cost here?
Opus 5 $3.25 / $16.25, Sonnet 5 $1.30 / $6.50, Haiku 4.5 $0.65 / $3.25 per 1M input / output tokens (Anthropic lists $5 / $25, $2 / $10, $1 / $5). Each model card above shows the live price.
Can I use Claude Code and the Anthropic SDK?
Yes. Set `ANTHROPIC_BASE_URL=https://api.crazyrouter.com` and use your Crazyrouter key as `ANTHROPIC_AUTH_TOKEN`. The Python and TypeScript SDKs accept the same `base_url` option.
Is there a free Claude API tier?
No — neither Anthropic nor Crazyrouter has a free tier. Crazyrouter is prepaid pay-as-you-go: add a small balance (card, Alipay or WeChat Pay) and you are billed per token at the rates shown for your model and request size.