Back

Anthropic

Claude Haiku 5.5

ReasoningClosed weights
Intelligence
37.8
Rank 82 of 744 measured models
Output speed
157 t/s
Rank 75 of 403 measured models
Web Search
$10 /1K calls
Context window
1M
Rank 84 of 1199 measured models
Cost per task
$0.0794
Rank 45 of 193 measured models

Claude Haiku 5.5 (High) scores 37.8 on the Artificial Analysis Intelligence Index, ranking 82nd of 744 measured models. The closest models on this index are Claude Sonnet 5 (38.2), GPT-6 Luna (38.1), and MiMo-V2.6-Flash (37.9). At $0.20 per million tokens (3:1 input/output blended price), it is cheaper than Claude Sonnet 5 ($4), priced close to GPT-6 Luna ($0.20), more expensive than MiMo-V2.6-Flash ($0.175). With an output speed of 157 t/s, it is faster than Claude Sonnet 5 (74.9 t/s), GPT-6 Luna (116 t/s), and MiMo-V2.6-Flash (53.8 t/s).

Artificial Analysis Indices
High
Intelligence37.8
Rank 82 of 744 measured models
Reasoning levels
Each point is a reasoning level of this model: more reasoning often raises the score but costs more. Select a point to open that level.
Intelligence Indexmost efficient ↗Cost per task
Capability indexes
Performance in specific industries, measured by Artificial Analysis
Finance & Accounting37.7
Rank 88 of 218 measured models
Strategy & Ops37.5
Rank 96 of 218 measured models
Legal37.6
Rank 89 of 218 measured models
Engineering39.2
Rank 72 of 217 measured models
Economics45.8
Rank 75 of 225 measured models
Decision Index
Claude Haiku 5.5 (High) · Artificial Analysis
Decision Index62.6Rank 7 of 26 measured models
AA-Omniscience-Decision · accuracy57.2
AA-Omniscience-Decision · non-hallucination48.6
HLE-MultiChoice26.7
When2Call76.9
WildGuardTest93.8
Cost per 1M tasks$292.17
Average time per task1.48 s
Decision Index leaderboard
Standard Benchmarks
HLE37.3%
Rank 105 of 664 measured models
SciCode48.7%
Rank 121 of 231 measured models
AA-LCR77.3%
Rank 130 of 575 measured models
Terminal-Bench v421.7%
Rank 56 of 218 measured models
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.

Preferences and analytics

Rybbit: visits, performance, clicks, copies, forms and session replay (masked inputs), with your permission. Refusing does not limit the site.

Privacy

Allow navigation analytics?

With your permission, Rybbit measures visits, performance and interactions (clicks, copies, forms), including session replay, to improve BenchSift. Input values are masked in replays. Refusing does not limit the site.

Your choice is stored for 180 days. Change it at any time using “Privacy preferences” in the footer. Withdrawal reloads the page.

No choice stored: Rybbit is disabled.

Privacy