Step 3.5 Flash
nvidia/stepfun-ai/step-3.5-flash · in: text → out: text · price provider-declared (via models.dev). Capabilities: tools reasoning text.
free
Blended $/1M (3:1)
free
Input $/1M
free
Output $/1M
free
1M in + 1M out
—
Cache read $/1M
#574 of 7,708
Cheapest rank
- Provider
- Nvidia (
nvidia) - Kind
- text (outputs text)
- Context
- 256,000 tokens
- Max output
- 16,384 tokens
- Capabilities
- tools reasoning text
Same model, other providers
“Step 3.5 Flash” is offered by 9 providers — cheapest-effective first. This one (Nvidia) is the cheapest.
| Provider | Input | Output | Blended | Context |
|---|---|---|---|---|
| Nvidia (this) | free | free | $0.00 | 256k |
| EmpirioLabs AI | $0.1 | $0.3 | $0.15 | 256k |
| NanoGPT | $0.1 | $0.3 | $0.15 | 262k |
| Kilo Gateway | $0.1 | $0.3 | $0.15 | 262k |
| OpenRouter | $0.1 | $0.3 | $0.15 | 262k |
| ZenMux | $0.1 | $0.3 | $0.15 | 256k |
| Hugging Face | $0.1 | $0.3 | $0.15 | 262k |
| StepFun (Global) | $0.1 | $0.3 | $0.15 | 256k |
| StepFun (China) | $0.1 | $0.3 | $0.15 | 256k |
Call it
jishie indexes & prices models; it does not proxy inference.
Most providers are OpenAI-compatible — point base_url at Nvidia and pass this model id:
curl $BASE_URL/v1/chat/completions \
-H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" \
-d '{"model":"stepfun-ai/step-3.5-flash","messages":[{"role":"user","content":"hello"}]}'
Route programmatically with MCP find_model / get_model, or fetch this record free (no wallet):