- Intelligence
- 28.5
- Rank 135 of 768 measured models
- Output speed
- 141 t/s
- Rank 113 of 425 measured models
- Price/1M
- $1.43
- Rank 461 of 726 measured models
- Context window
- 1M
- Rank 90 of 1296 measured models
- Cost per task
- $2.35
- Rank 172 of 218 measured models
Step 5 Preview scores 28.5 on the Artificial Analysis Intelligence Index, ranking 135th of 768 measured models. The closest models on this index are DeepSeek V4.1 Flash (28.6), MiMo-V2-Pro (28.6), and GPT-5.2 Codex (Xhigh) (28.5). At $1.43 per million tokens (3:1 input/output blended price), it is cheaper than GPT-5.2 Codex (Xhigh) ($4.81), more expensive than DeepSeek V4.1 Flash ($0.525). With an output speed of 141 t/s, it is faster than GPT-5.2 Codex (Xhigh) (99.9 t/s), slower than DeepSeek V4.1 Flash (203 t/s).
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.
- DeepSeek V4.1 Flash
DeepSeek · $0.525 · 203 t/s
28.6+0.1$0.525203 t/sCompare - MiMo-V2-Pro
Xiaomi · — · —
28.6+0.1——Compare - GPT-5.2 Codex (Xhigh)
OpenAI · $4.81 · 99.9 t/s
28.5$4.8199.9 t/sCompare - Qwen3.6 Max Preview
Alibaba · $2.93 · 62.3 t/s
28.4−0.1$2.9362.3 t/sCompare
