Step 3.7 Flash
AI model by StepFun. Real-time pricing and benchmark data.
Benchmarks
Editorial Analysis
Profile. Step 3.7 Flash is a budget-tier option from StepFun at $1.15/M output tokens. Despite the low price, it scores competitively on Coding Index (39.6) and Intelligence Index (30.9) — making it a candidate for high-volume production workloads where unit cost dominates.
Latency. On latency, Step 3.7 Flash reports 106 tok/s output speed and 2730ms time-to-first-token. Output speed is in the normal range — most chat and completion applications will be token-bound, not streaming-bound.
Cost model. For a workload of 1M input tokens + 100K output tokens, Step 3.7 Flash costs $0.32 per million input-equivalent requests. Multiply by your monthly request volume to project the bill, and compare against the speed/correctness profile above to decide if the trade-off is worth it for your workload.
Editorial summary generated from public pricing and benchmark data (Artificial Analysis API). All numbers are sourced — see the Artificial Analysis methodology for benchmark definitions. Last refreshed 2026-09-06.
Compare with similar models
| Model | Input | Output | Speed |
|---|---|---|---|
Step 3.7 FlashCurrent | $0.20 | $1.15 | 106 tok/s |
| DeepSeek V3 0324 | $0.27 | $1.12 | — |
| DeepSeek R1 Distill Llama 70B | $0.70 | $1.10 | 28 tok/s |
| ERNIE 4.5 300B A47B | $0.28 | $1.10 | — |
| GPT-5.6 Luna (low) | $0.20 | $1.20 | 109 tok/s |
| MiniMax-M2.5 | $0.30 | $1.20 | 97 tok/s |