StepFun: Step 3.7 Flash
by StepFunStepFun: Step 3.7 Flash is a large language model from StepFun. It costs $0.200 per million input tokens and $1.15 per million output tokens. Its context window is 262K tokens.
- Input / 1M tokens
- $0.200
- Output / 1M tokens
- $1.15
- Cached input / 1M
- $0.040
- Context window
- 262K
On repeated prefixes
tokens
Who serves it cheapest
3 hosts serve Step 3.7 Flash. Same weights, same API - the price difference is pure margin and routing.
| Provider | Input / 1M | Output / 1M | Context | Throughput | Uptime 24h |
|---|---|---|---|---|---|
| DeepInfraCheapest | $0.160 | $0.920 | 262K | - | 100.0% |
| Novitafp8 | $0.200 | $1.15 | 262K | - | 99.3% |
| StepFunfp8 | $0.200 | $1.15 | 256K | - | 99.4% |
The spread between DeepInfra and StepFun is 1.3× for identical weights. Quantization and context limits differ, so check both columns before switching.
Benchmarks
Independent scores published alongside the catalogue.
- Coding index
- 39.6
About Step 3.7 Flash
Step 3.7 Flash is StepFun's latest high-efficiency multimodal Mixture-of-Experts model. It pairs a 196B-parameter language backbone with a vision encoder for native image and video understanding, activating roughly 11B parameters...
Specifications
| Model ID | stepfun/step-3.7-flash |
|---|---|
| Provider | StepFun |
| Context window | 262K tokens |
| Max output | 230K tokens |
| Input modalities | text, image, video |
| Output modalities | text |
| Knowledge cutoff | - |
| Open weights | Yes - stepfun-ai/Step-3.7-Flash |
| Released | May 28, 2026 |
Cheaper alternatives
Models that cost less than Step 3.7 Flash while keeping at least half its context window and every input modality it supports.
Frequently asked
How much does StepFun: Step 3.7 Flash cost?
$0.200 per million input tokens and $1.15 per million output tokens. Cached input reads cost $0.040 per million tokens.
What is the context window of StepFun: Step 3.7 Flash?
262K tokens, with up to 230K tokens of output per request.
Which provider serves StepFun: Step 3.7 Flash cheapest?
DeepInfra at $0.160 per million input tokens - 1.3× cheaper than StepFun, the most expensive of the 3 hosts serving it.