Skip to content

We raised $7.5m to build the intelligence layer for AI

ModelsPricing Docs

Sign in Start building

Features

Spend controlsBudgets and limits for every key and teammate
GuardrailsBlock risky requests before a model sees them.
Usage analyticsSpend, speed and failures across your requests
Request logsEvery request with its cost and outcome
FallbacksBackup models, tried in the order you set
Bring your own keyYour provider keys, with limits and tracking
PlaygroundTest prompts on up to three models
IntegrationsCoding agents, SDKs and frameworks

Use cases

Agent buildersComing soon
AI startupsComing soon
AgenciesComing soon
Enterprise teamsComing soon

Resources

BlogComing soon
Case studiesComing soon
DiscordAsk the team and other developers

Qwen3.8 Flash

deepinfra/Qwen3.8-Flash

Released Aug 26, 2026

Providers: 1
Context: —
Max out: —
From / 1M: $0.113

Capabilities

What the model accepts and returns.

Tools
JSON mode
Vision
Audio
Streaming

Providers

Cheapest first.

1 of 11 providers

DeepInfraQwen/Qwen3.8-Flash
deepinfra/Qwen3.8-Flash
Input / 1M
$0.113
Output / 1M
$0.382
Cache read / 1M
N/A

Hopscotch is an LLM gateway: one API for top models, with spend limits and failover built in.

© 2026 Hopscotch

Product

Spend controls
Guardrails
Usage analytics
Request logs
Fallbacks
Bring your own key
Playground
Integrations
Pricing

Models

All models
OpenAI
Google
Alibaba
Zhipu AI
Anthropic
DeepSeek
Meta
MiniMax

Developers

Docs
Quickstart
API reference
llms.txt

Company

Team
Contact
Discord

Legal

Privacy Policy
Terms of Service
Service Level Agreement

AnthropicCohereDeepInfraDeepSeekFireworks AIGoogleMetaMiniMaxMoonshot AIOpenAIZ.ai

hopscotch