Haiku 4.5 Benchmark Details
Haiku 4.5 currently shows benchmark results led by AIME2025 (20 / 106, score 96.30), Terminal-Bench (11 / 35, score 41), Claw Bench (11 / 29, score 89.40). This page also compares it with 2 competitor models and 2 predecessor or same-series models, including performance and pricing views when available.
Benchmark Results
Benchmark Results
General Knowledge
12 evaluationsCoding and Software Engineer
5 evaluationsMath and Reasoning
5 evaluationsAI Agent - Tool Usage
3 evaluationsClaw-style Agent Evaluation
2 evaluationsCompetitor Comparison
Benchmark scores for Haiku 4.5 compared against top models in its class
11 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.
| Benchmark | Haiku 4.5Current | GPT-5.4 mini | Gemini 3.0 Flash |
|---|---|---|---|
ARC-AGI-2 综合评估 | 4.50Extended Thinking | -- | 33.60Thinking Enabled |
GPQA Diamond 综合评估 | 73.30Extended Thinking | 88.00Thinking Level · Extra High | 90.40Thinking Enabled |
HLE 综合评估 | 9.70Extended Thinking | 41.50Thinking Level · Extra High ｜ Tools | 43.50Thinking Enabled ｜ Tools |
LiveBench 综合评估 | 61.3264K | 67.54Deep Thinking Mode | 72.40Thinking Level · High |
SWE-Bench Pro - Public 编程与软件工程 | 39.45Extended Thinking ｜ Tools | 54.40Thinking Level · Extra High ｜ Tools | 49.60Thinking Level · High ｜ Tools |
SWE-bench Verified 编程与软件工程 | 73.30128K ｜ Tools | -- | 68.70Thinking Enabled |
AIME2025 数学推理 | 96.30128K ｜ Tools | -- | 99.70Thinking Enabled ｜ Tools |
2.1032K | 2.10Thinking Level · High | 4.20Standard Mode | |
τ²-Bench Agent能力评测 | 33.00Standard Mode ｜ Tools | -- | 90.20Thinking Enabled ｜ Tools |
Claw Bench OpenClaw智能体能力综合测评 | 89.40Thinking Enabled ｜ Tools | 75.30Thinking Enabled ｜ Tools | 85.70Thinking Enabled ｜ Tools |
Pinch Bench OpenClaw智能体能力综合测评 | 82.00Thinking Enabled ｜ Tools | -- | 85.20Thinking Enabled ｜ Tools |
Standard API Pricing: Haiku 4.5 vs. Peer Models
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier. · USD / 1M tokens
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
Haiku 4.5 | Anthropic | $1 / 1M tokens | $5 / 1M tokens | — |
GPT-5.4 mini | OpenAI | $0.75 / 1M tokens | $4.5 / 1M tokens | — |
Gemini 3.0 Flash | Google Deep Mind | $0.5 / 1M tokens | $3 / 1M tokens | — |
Version History
How each version of the Haiku 4.5 series stacks up on benchmark tests
2 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.· Click a row to view its trend chart.
| Benchmark | Haiku 4.5Current | Claude 3.5 Haiku |
|---|---|---|
GPQA Diamond 综合评估 | 73.30Extended Thinking | 41.60Standard Mode |
MMLU Pro 综合评估 | 80.00Extended Thinking | 65.00Standard Mode |
Single-Benchmark Version Trend
Viewing: GPQA Diamond · 综合评估
Standard API Pricing Across the Haiku 4.5 Series
Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.
Source: DataLearnerAI. Standard text prices shown here use the default supplier. · USD / 1M tokens
| Model | Supplier | Standard input | Standard output | Base price applies to |
|---|---|---|---|---|
Haiku 4.5 | Anthropic | $1 / 1M tokens | $5 / 1M tokens | — |
Claude 3.5 Haiku | Anthropic | $0.8 / 1M tokens | $4 / 1M tokens | — |
Claude3-Haiku | Anthropic | $0.25 / 1M tokens | $1.25 / 1M tokens | — |