DataLearner logo
HA

Haiku 4.5

Multimodal modelHaikuClaude 4.5

Claude Haiku 4.5

Release date: 2025-10-15Updated: 2026-06-15 07:18:22.2731,510
Live demoGitHubHugging FaceCompare
Parameters
Not disclosed
Context length
200K
Chinese support
Supported
Reasoning ability

Claude Haiku 4.5 is an AI model published by Anthropic, released on 2025-10-15, for Multimodal model, and 200K context length, under the 不开源 license, with a 96.30 score on AIME2025.

Data sourced primarily from official releases (GitHub, Hugging Face, papers), then benchmark leaderboards, then third-party evaluators. Learn about our data methodology

Haiku 4.5

Model basics

Reasoning traces
Supported
Thinking modes
Thinking modes not supported
Context length
200K tokens
Max output length
64K tokens
Model type
Multimodal model
Modality (in / out)
Text, Image → Text
Release date
2025-10-15
Model file size
No data
MoE architecture
No
Total params / Active params
No data / N/A
Knowledge cutoff
No data
Haiku 4.5

Open source & experience

Code license
不开源
Weights license
不开源
GitHub repo
GitHub link unavailable
Hugging Face
Hugging Face link unavailable
Haiku 4.5

Official resources

Haiku 4.5

API details

API speed
5/5
💡Default unit: $/1M tokens. If vendors use other units, follow their published pricing.
Standard
TypeConditionInputOutput
Text-$1.00/ 1M$5.00/ 1M
Batch
TypeConditionInputOutput
Text-$0.500/ 1M$2.50/ 1M
Cache PricingPrompt Cache
TypeTTLWriteRead
Text-$1.25/ 1M$0.100/ 1M
Haiku 4.5

Benchmark Results

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 consolidates core specs, context limits, and API pricing so you can evaluate the model from benchmark results and deployment constraints together.

Thinking
Tool usage

General Knowledge

12 evaluations
Benchmark / mode
Score
Rank/total
MMLU Pro
Standard Mode
76
78 / 126
MMLU Pro
Extended
80
60 / 126
GPQA Diamond
Standard Mode
60.50
140 / 180
GPQA Diamond
Extended
73.30
102 / 180
LiveBench
Standard Mode
45.33
103 / 115
61.32
64 / 115
ARC-AGI
Standard Mode
14.30
56 / 65
ARC-AGI
Extended
47.70
43 / 65
HLE
Standard Mode
4.30
162 / 164
HLE
Extended
9.70
140 / 164
ARC-AGI-2
Standard Mode
1.30
52 / 59
ARC-AGI-2
Extended
4.50
47 / 59

Coding and Software Engineer

5 evaluations
Benchmark / mode
Score
Rank/total
SWE-bench Verified
Standard ModeTools
60.60
77 / 109
73.30
45 / 109
LiveCodeBench
Standard Mode
51
91 / 120
62
67 / 120
39.45
44 / 47

Math and Reasoning

5 evaluations
Benchmark / mode
Score
Rank/total
AIME2025
Standard Mode
39
94 / 106
80.70
57 / 106
AIME2025
128KTools
96.30
20 / 106
FrontierMath
Standard Mode
4.10
41 / 60
2.10
56 / 80

AI Agent - Tool Usage

3 evaluations
Benchmark / mode
Score
Rank/total
Terminal-Bench
Standard ModeTools
26
26 / 35
41
11 / 35
MCP-Atlas
Standard ModeTools
40.20
25 / 25

Multimodal Understanding

1 evaluations
Benchmark / mode
Score
Rank/total
MMMU
128K
73.20
19 / 28

Agent Level Benchmark

1 evaluations
Benchmark / mode
Score
Rank/total
τ²-Bench
Standard ModeTools
33
40 / 40

Instruction Following

1 evaluations
Benchmark / mode
Score
Rank/total
IF Bench
Extended
54.30
25 / 29

Claw-style Agent Evaluation

2 evaluations
Benchmark / mode
Score
Rank/total
Claw Bench
Thinking ModeTools
89.40
11 / 29
Pinch Bench
Thinking ModeTools
82
21 / 37

Compare with other models

Haiku 4.5

Publisher

Claude Haiku 4.5

Model Overview

Claude Haiku 4.5 is an AI model published by Anthropic, released on 2025-10-15, for Multimodal model, and 200K context length, under the 不开源 license, with a 96.30 score on AIME2025.

DataLearner on WeChat

Follow DataLearner on WeChat for AI model updates and research notes.

DataLearner WeChat QR code