Anthropic
Claude 4.5 Haiku
ReasoningClosed weights
- Intelligence
- 15.4
- Rank 304 of 744 measured models
- Coding
- 43.9
- Rank 144 of 298 measured models
- Output speed
- 87.3 t/s
- Rank 204 of 403 measured models
- Price/1M
- $2
- Rank 466 of 659 measured models
- Context window
- 200K
- Rank 618 of 1199 measured models
Claude 4.5 Haiku (Non-reasoning) scores 15.4 on the Artificial Analysis Intelligence Index, ranking 304th of 744 measured models. The closest models on this index are Gemini 3.1 Flash-Lite (15.6), Claude 3.7 Sonnet (15.3), and Kimi K2 0905 (15.3). At $2 per million tokens (3:1 input/output blended price), it is cheaper than Claude 3.7 Sonnet ($6), more expensive than Gemini 3.1 Flash-Lite ($0.563) and Kimi K2 0905 ($1.08). With an output speed of 87.3 t/s, it is faster than Kimi K2 0905 (56.1 t/s), slower than Gemini 3.1 Flash-Lite (342 t/s).
Additional benchmarks8
Design Arena · 3d41.1%
Design Arena · Asciiart49.3%
Design Arena · Codecategories44.9%
Design Arena · Dataviz45.6%
Design Arena · Gamedev44.7%
Design Arena · Svg39.1%
Design Arena · Uicomponent42.6%
Design Arena · Website45.1%
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.
- Gemini 3.1 Flash-Lite
Google · $0.563 · 342 t/s
15.6+0.2$0.563342 t/sCompare - K2 Horizon 3.7B
Institute of Foundation Models · — · —
15.6+0.2——Compare - Claude 3.7 Sonnet
Anthropic · $6 · —
15.3−0.1$6—Compare - Kimi K2 0905
Moonshot AI · $1.08 · 56.1 t/s
15.3−0.1$1.0856.1 t/sCompare
