Anthropic: API rate limits raised — Sonnet and Haiku now match Opus across three tiers
Anthropic has equalized API rate limits across all models — Sonnet and Haiku now share the same quotas as Opus at each of the three usage tiers (Start, Build, Scale). At the same time, fast mode for Claude Opus 4.7 is being deprecated, with removal scheduled for July 24.
This article was generated using artificial intelligence from primary sources.
What are rate limits and why does this change matter?
Rate limits — the upper bound on API requests and tokens an organization can send within a given time window — have long favored Opus models. Developer teams using Sonnet or Haiku (the cheaper, faster Claude variants) automatically had narrower quotas than Opus users, limiting the scalability of production systems. Anthropic changed that in a single move: Sonnet and Haiku now have identical limits to Opus at every tier.
What are the new tiers and who moves to them?
Anthropic has consolidated the structure into three usage tiers: Start, Build, and Scale. The previous system used a different breakdown, but migration is automatic — organizations move to an equivalent or higher tier with no action required on their part. No organization receives lower limits than before. Current quotas are visible in Claude Console under organization settings.
Opus 4.7 fast mode is going, Opus 4.8 takes over
Alongside the tier change, Anthropic announced on June 25 the deprecation of fast mode for Claude Opus 4.7 — removal is scheduled for July 24, 2026. Fast mode, which enables faster generation with less emphasis on depth of reasoning, will move exclusively to Claude Opus 4.8, the newer generation model with an improved speed-to-quality ratio. Developer teams currently using claude-opus-4-7 with fast mode enabled need to update their calls to claude-opus-4-8 before the deadline — the API will still accept requests, but without fast mode, which may increase latency in production pipelines.
Practical implication
For Sonnet and Haiku users this is an immediate gain: higher throughput without upgrading a plan. For Opus 4.7 fast mode users, the migration window is just under four weeks.
Frequently Asked Questions
- What are usage tiers and what does the move to three tiers mean?
- A usage tier is a level of API access that determines the maximum number of requests and tokens per minute. Anthropic has consolidated its offering into three tiers — Start, Build, and Scale — with no organization receiving lower limits than before.
- Do I need to take action due to the Opus 4.7 fast mode deprecation?
- Yes — by July 24, 2026, you need to migrate any calls using Opus 4.7 fast mode to Claude Opus 4.8 fast mode, which takes over that role without any loss of functionality.