Primary use stated or inferred cautiously from official documentation.
A
ANTHROPIC / claude-haiku-4-5-20251001
Claude Haiku 4.5
High volume, subagents, and structured extraction.
01 / SUMMARY
The essential
Maximum input capacity when published by the source.
Documented maximum generation limit.
Verified availability channels.
Primary functional family and verified specialties.
Declared terms for API access, model weights, or self-hosting.
02 / TECHNICAL SHEET
Limits and integration
| API ID | claude-haiku-4-5-20251001 |
|---|---|
| Model type | Text and reasoning |
| Access model | Paid proprietary |
| License | Proprietary |
| Deployment | Hosted API |
| Release | 15 OCT 2025 |
| Knowledge cutoff | FEB 2025 |
| Entrance | Text · Image |
| Exit | Text |
| Context window | 200.000 tokens |
| maximum output | 64.000 tokens |
| Reasoning | Extended thinking |
| Published tools | Use of tools · Web search · Code execution · Computer use |
| Structured Outings | Supported |
| Batch processing | Compatible · 50% discount |
| Prompt cache | Compatible · up to 90% savings on reads |
| Fine-tuning | Not published |
| Verified platforms | Claude.ai · Claude API · Amazon Bedrock · Google Cloud · Microsoft Foundry |
High volume, subagents, and structured extraction.
Cost and performance depend on the reasoning level.
03 / CAPABILITIES
What it can do
Orientation
High volume, subagents, and structured extraction.
Context
200.000 tokens · 64.000 tokens
Tools and integration
Tool use · Web search · Code execution · Computer use
Access
Claude, Claude API and associated clouds
Input modalities
TextImage
04 / PRICES
Documented cost
| Concept | Worth | Unit/condition |
|---|---|---|
| Standard input | 1.00 USD / 1 M tokens | Standard API rate |
| Cached input | 0,10 USD / 1 M tokens | Reading reused prefixes |
| Cache write or storage | 1,25–2,00 USD / 1 M tokens | The condition varies by provider |
| Standard output | 5.00 USD / 1 M tokens | May include reasoning tokens |
| Batch input | 0.50 USD / 1 M tokens | Asynchronous processing |
| Batch output | 2.50 USD / 1 M tokens | Asynchronous processing |
Consult the primary source before budgeting for a deployment.
i Prices change and may depend on level, region or context length. Check the source before making a decision.
05 / EVALUATIONS
How to read the results
“
A public benchmark provides guidance, but does not replace an evaluation with your data, tools, budget, and error tolerance.
| Benchmark | Result | Metric | Source |
|---|---|---|---|
| SWE-bench Verified | 73,3 % | Resolved | View source ↗ |
| Terminal-Bench | 41,75 % | With reasoning | View source ↗ |
Inferama only highlights a “best result” when the metric, test set, configuration, and date allow for an equivalent comparison. The supplier's figures are presented as claims from its own source.
06 / VERSIONS
Chronology
Claude Haiku 4.5
Version added to Inferama's verified catalog.
ANALYSIS
Related analysis
Claude Haiku 4.5: How to Calculate Cost per Usable Response
The price per call does not, by itself, show how much an accepted response costs. This guide provides a reproducible formula for adding input and output tokens, retries, and review, with three illustrative scenarios for Claude Haiku 4.5.
23 Sep 2026 ↗
Claude Haiku 4.5: How to Operate a Low-Latency Lane After Haiku 3.5 Without Mistaking Minimum Retention for Guaranteed Continuity
Claude Haiku 4.5 can be a destination for fast workloads after Claude Haiku 3.5 was retired, but a safe migration is not solved by changing an identifier. Documented minimum availability through October 15, 2026 does not guarantee indefinite continuity. The operational criterion is to demonstrate, using your own traffic and corpus, queue latency, schema compliance, tool use, and a reversible replacement path.
23 Sep 2026 ↗
Cost in SWE-bench: how to calculate the price per resolved issue without hiding failures, retries, or evaluation
A cost figure per resolved issue is useful only if it reveals everything that happened before the patch was obtained: failed attempts, token consumption, stopping rules, selection among runs, and evaluation resources. This analysis proposes a reproducible scorecard for reading and comparing SWE-bench results, with particular attention to SWE-bench Verified.
22 Sep 2026 ↗
Analyzing customer feedback with AI: how to turn thousands of comments into priorities without confusing frequency with impact
A path for product, support, and research teams that need to analyze reviews, tickets, surveys, or transcripts with AI without turning an automated summary into an unsupported decision.
22 Sep 2026 ↗
Claude Haiku 4.5 vs Claude Opus 5: when a cost premium can justify a better result
The choice between Claude Haiku 4.5 and Claude Opus 5 should not be settled solely by per-token pricing or unrelated benchmarks. This comparison proposes a reproducible protocol for measuring accepted outputs, retries, human review, latency, and effective cost in structured classification, document synthesis, and complex technical review. The available sources can establish comparable conditions and technical constraints, but they cannot support empirical results without running and publishing the experiment.
22 Sep 2026 ↗
Anthropic: how to verify what changes when you use Claude through an API, partner cloud, or product
Adopting Claude is not simply a matter of selecting a model family. The access channel determines which identifier is used, which retirement schedule governs it, which controls each party administers, and which documentation can support a technical decision. This guide offers a method for separating those layers without turning public policies or safety evaluations into guarantees they do not contain.
22 Sep 2026 ↗07 / SOURCES