Is Claude Haiku 3.5 deprecated?
Anthropic · lifecycle status verified Jul 25, 2026 against the provider's own notice.
Last verified pricing
historical reference — not a buy recommendationClaude Haiku 3.5 was last priced at $0.800 per 1M input tokens and $4.00 per 1M output tokens, with a 200K-token context window. Its benchmark average was 78.8 on the blend we track. Full history stays on the Claude Haiku 3.5 model page.
What replaced Claude Haiku 3.5?
successor named by AnthropicAnthropic's named successorfrom the deprecation notice — not our inference
| Model | Input $/M | Output $/M | Context | |
|---|---|---|---|---|
| Claude Haiku 3.5 Retired | $0.800 | $4.00 | 200K | |
| Claude Haiku 4.5 | $1.00 | $5.00 | 200K | details → |
Same benchmark score or better, for less
computed from live prices + benchmark averagesCheaper equivalentsscore ≥ 78.8 on the benchmark blend, ranked by blended cost ((3×input + 1×output) ÷ 4)
| Model | Benchmark | Blended $/M | Input $/M | Output $/M | |
|---|---|---|---|---|---|
| Claude Haiku 3.5 Retired | 78.8 | $1.60 | $0.800 | $4.00 | |
| DeepSeek V4 Flash DeepInfra cheapest | 82.2 | $0.110 | $0.090 | $0.180 | details → |
| GPT-OSS 20B Groq best value | 98.7 | $0.130 | $0.075 | $0.300 | details → |
"Equivalent" here means the model's benchmark average is at or above Claude Haiku 3.5's 78.8. Benchmarks compress real quality differences — validate on your own workload before switching. Model your exact token mix in the migration calculator.
Current Anthropic alternatives
same provider, sold today — no parity claim| Model | Input $/M | Output $/M | Context | |
|---|---|---|---|---|
| Claude Haiku 4.5 | $1.00 | $5.00 | 200K | details → |
Migrating off Claude Haiku 3.5? The migration calculator compares your exact input/output token mix against every current model we track, with benchmark tradeoffs. Lifecycle data on this page comes from Anthropic's own notice and was last re-verified on Jul 25, 2026. See all retired & deprecated models.