Claude Haiku 4.5
- Context
- 200K
- 64K max output
- Input price
- $1.00
- per 1M tokens · Claude API
- Output price
- $5.00
- per 1M tokens
- Epoch Capabilities Index
- 142.4
- #67 of 71 listed
- LMArena Text
- 1,414
- #62 of 79 listed
Where to run it
5 offerings · prices as listed| Meter | Tier | Price |
|---|---|---|
| Input | Standard | $1.00 / 1M tokens |
| Batch | $0.50 / 1M tokens | |
| Output | Standard | $5.00 / 1M tokens |
| Batch | $2.50 / 1M tokens | |
| Cache read | Standard | $0.10 / 1M tokens |
| Cache write | 5-minute cache | $1.25 / 1M tokens |
| 1-hour cache | $2.00 / 1M tokens | |
Amazon Bedrock (opens in a new tab) Claude Haiku 4.5 | ||
| Input | Standard | $1.00 / 1M tokens |
| Batch | $0.50 / 1M tokens | |
| Output | Standard | $5.00 / 1M tokens |
| Batch | $2.50 / 1M tokens | |
| Cache read | Standard | $0.10 / 1M tokens |
| Cache write | 5-minute cache | $1.25 / 1M tokens |
| 1-hour cache | $2.00 / 1M tokens | |
Google Cloud Vertex AI (opens in a new tab) claude-haiku-4-5 | ||
| Input | Standard | $1.00 / 1M tokens |
| Batch | $0.50 / 1M tokens | |
| Output | Standard | $5.00 / 1M tokens |
| Batch | $2.50 / 1M tokens | |
| Cache read | Standard | $0.10 / 1M tokens |
| Batch | $0.050 / 1M tokens | |
| Cache write | 5-minute cache | $1.25 / 1M tokens |
| 1-hour cache | $2.00 / 1M tokens | |
| 5-minute cache · Batch | $0.63 / 1M tokens | |
Google Cloud Vertex AI (opens in a new tab) Claude Haiku 4.5 | ||
| Input | Standard | $1.10 / 1M tokens |
| Batch | $0.55 / 1M tokens | |
| Output | Standard | $5.50 / 1M tokens |
| Batch | $2.75 / 1M tokens | |
| Cache read | Standard | $0.11 / 1M tokens |
| Batch | $0.055 / 1M tokens | |
| Cache write | Standard | $1.38 / 1M tokens |
| 5-minute cache | $1.38 / 1M tokens | |
| 1-hour cache | $2.20 / 1M tokens | |
| 5-minute cache · Batch | $0.69 / 1M tokens | |
| 1-hour cache · Batch | $1.10 / 1M tokens | |
Google Cloud Vertex AI (opens in a new tab) Claude Haiku 4.5 | ||
| Input | Standard | $1.10 / 1M tokens |
| Batch | $0.55 / 1M tokens | |
| Output | Standard | $5.50 / 1M tokens |
| Batch | $2.75 / 1M tokens | |
| Cache read | Standard | $0.11 / 1M tokens |
| Batch | $0.055 / 1M tokens | |
| Cache write | 5-minute cache | $1.38 / 1M tokens |
| 1-hour cache | $2.20 / 1M tokens | |
| 5-minute cache · Batch | $0.69 / 1M tokens | |
| 1-hour cache · Batch | $1.10 / 1M tokens | |
Changes
All changes →- 3 new prices on Claude API Cache write (1-hour cache) $2.00 · Input (Batch) $0.50 · Output (Batch) $2.50
Vendor-reported scores
Published by Anthropic, under the conditions it states| Benchmark | Setting | Score |
|---|---|---|
| SWE-bench Verified | bash + string-replace edit tools, averaged over 50 trials, 128K thinking budget, prompt addition encouraging tool use | 73.3% |
| τ²-bench | Retail domain; extended thinking (128K budget), averaged over 10 runs, prompt addendum | 83.2% |
| Airline domain; extended thinking (128K budget), averaged over 10 runs, prompt addendum | 63.6% | |
| Telecom domain; extended thinking (128K budget), averaged over 10 runs, prompt addendum | 83% | |
| OSWorld-Verified | OSWorld-Verified framework, 100 max steps, averaged over 4 runs, 128K total / 2K per-step thinking budget | 50.7% |
| AIME 2025 | with Python; pass@1 over 16 trials averaged over 10 runs, 128K thinking budget | 96.3% |
| no tools; pass@1 over 16 trials averaged over 10 runs, 128K thinking budget | 80.7% | |
| GPQA Diamond | averaged over 10 runs, 128K thinking budget | 73% |
| MMMLU | average of 10 runs over 14 non-English languages, 128K thinking budget | 83% |
| MMMU | validation set, averaged over 10 runs, 128K thinking budget | 73.2% |
Cited scores
From third-party leaderboards, not measured by us- Epoch Capabilities Index 142.4
#67 of 71 listed models · CI 139.4–144.1
Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Sep 26, 2026 · changes we made
- LMArena Text 1,414
#62 of 79 listed models · CI 1,412–1,416 · 140,791 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- Epoch AI · GPQA Diamond 71.2%
#61 of 64 listed models · 32K thinking budget · ±3.23
Also listed: claude-haiku-4-5-20251001 60.5%
Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Sep 26, 2026 · changes we made
- LMArena Text · Chinese 1,431
#67 of 79 listed models · CI 1,424–1,438 · 9,882 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made
- LMArena Text · Coding 1,482
#55 of 79 listed models · CI 1,478–1,486 · 36,489 votes
LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Sep 25, 2026 · changes we made