Pyyan / Compare / Step 5 Preview vs Claude Opus 5.5
Step 5 Preview vs Claude Opus 5.5
2 of 5Language Models · verified 24 Sept 2026
Step 5 PreviewStepFuncurrentClaude Opus 5.5Anthropiccurrent
| Specification | Step 5 Preview | Claude Opus 5.5 |
|---|---|---|
| Summary | A frontier tier score at a tenth of the frontier price. | Fable 5.1's level of work on most tasks, at 40% less to run than Opus 5. |
| Context | 1M | 1M |
| Input ($/Mtok) | $1 | $4 |
| Output ($/Mtok) | $2.7 | $20 |
| Licence | Proprietary | Proprietary |
| Max output | Not published | 128K |
| Parameters | 600B total, 27B active | Not published |
| Released | 18 Sep 2026 | 22 Sep 2026 |
| Model ID | step-5-preview | claude-opus-5-5 |
| Category | Language Models | Language Models |
| Official | StepFun ↗ | Anthropic ↗ |
Highlighted rows are where these differ.
Step 5 Preview
- 600B total parameters with 27B active per token, text and images in
- 24th of 200 models on the Artificial Analysis intelligence index
- Listings disagree on context: 1M on StepFun's API and Artificial Analysis, 100K on some trackers
Best for cheap reasoning at volume.
Claude Opus 5.5
- 66.4% on Terminal-Bench 4.0 against 52.3% for Opus 5, on Anthropic's own harness
- Cache reads cut 60% to $0.20, and output more than 30% faster than Opus 5
- Attempts to work around its boundaries 85% less often than Opus 5, Anthropic reports
Best for agentic coding at flagship quality.