Compare Models

Select models (max 7)
Claude Haiku 4.5 (Thinking)Hy4
Benchmarks

Vals Index *

Claude Haiku 4.5 (Thinking)
N/A
Hy4
0.00%± 1.08
(33/33)

Legal Research Bench *

Claude Haiku 4.5 (Thinking)
0.00%± 2.14
(90/90)
Hy4
0.00%± 3.42
(90/90)

Finance Agent v2.1 *

Claude Haiku 4.5 (Thinking)
N/A
Hy4
0.00%± 2.05
(36/36)

Tax Agent Bench *

Claude Haiku 4.5 (Thinking)
0.00%± 1.54
(81/81)
Hy4
0.00%± 3.30
(81/81)

MedCode *

Claude Haiku 4.5 (Thinking)
0.00%± 2.00
(107/107)
Hy4
N/A

Terminal-Bench 4.0

Claude Haiku 4.5 (Thinking)
N/A
Hy4
0.00%± 0.88
(60/60)

Vibe Code Bench v1.1 *

Claude Haiku 4.5 (Thinking)
0.00%± 3.13
(124/124)
Hy4
0.00%± 3.55
(124/124)

Overall performance

Performance on the Vals Index, a GDP-weighted aggregation of tasks across finance, coding, legal, and tax

2026-11-04

Comparison by Industry

Model performance on different sections of the economy.

Legal
45.91%
26.92%
Finance
45.19%
55.09%
Healthcare
58.95%
N/A
Academic
65.67%
N/A
Coding
29.79%
60.20%

Cost Analysis

Per task and per token model pricing.

Vals Index · USD
N/A
$1.72
Cost / TestVals Index
N/A
$1.72
Input Cost/ 1M Tokens
$1.00
$0.834
Input Cache Write/ 1M Tokens
$1.25
N/A
Input Cache Read/ 1M Tokens
$0.10
$0.042
Output Cost/ 1M Tokens
$5.00
$2.501

Model Metadata

Basic information about each model.

Model provider
AnthropicAnthropic
TencentOpenrouter
LatencyN/A3091.47s
Cost (In/Out)$1 / $5$0.834 / $2.501
Context Window200k1M
Max Output Token64k66k
Input Modality