The AI arena is free today

Open Superagent
API Provider102 active models22 organizations

DeepInfra: API pricing, speed & models

DeepInfra hosts 102 active AI models, with input pricing from $0.02 per 1M tokens, with median throughput of 56 characters/sec, and P95 time to first token of 4.26s, with 50% success rate over 7 days. Compare DeepInfra's API speed, pricing, and reliability against other inference providers.

Median throughput56 c/s
Median TTFT1.09s
P95 TTFT4.26s
Success rate (7d)50%
From$0.02 /M tok

Catalog

Alibaba Cloud / Qwen Team39Google37Anthropic14DeepSeek14ByteDance12Meta11NVIDIA10Moonshot AI8
Type
Price
192 models
Model
Qwen3.8 Maxfp4cache pricing
Qwen3.8-27Bcache pricing
Qwen3.8-27Bcache pricing
Qwen3.7 Maxcache pricing
Qwen3.5-35B-A3Bfp8cache pricing
Qwen3.5-35B-A3Bfp8cache pricing
Qwen3.5-35B-A3Bfp8cache pricing
Qwen3 MaxExclusivecache pricing
Qwen3.8 Maxcache pricing
At a glance

DeepInfrapricing, performance & catalog

The citable facts about DeepInfra's 102 models — sourced from provider APIs and refreshed continuously.

Lowest price
Mistral NeMo Instruct at $0.019 per 1M input tokens
Highest median throughput
Kimi K2.6 at 69 chars/s
Lowest median TTFT
Kimi K2.6 at 1.06s
Largest context
DeepSeek-V4.1-Flash at 1.0M tokens
Catalog
102 active models from 22 organizations

Fastest median throughput

  1. 1Kimi K2.669 c/s
  2. 2Seed 2.0 Pro5 c/s

FAQ

Common questions about DeepInfra.

What is DeepInfra?

DeepInfra is an API provider that hosts large language models. Active models: 102; From (input): $0.02 / 1M tok; Median throughput: 56 c/s; Median TTFT: 1.09s; Success rate (7d): 50%.

How many models does DeepInfra offer?

DeepInfra currently serves 102 active models out of 123 historical offerings on LLM Stats.

What is DeepInfra's API pricing?

DeepInfra input pricing starts from $0.02 per 1M tokens, with the most expensive offering at $10 per 1M tokens. See the Pricing tab above for the full per-model breakdown.

How fast is DeepInfra?

DeepInfra delivers a median throughput of 56 characters per second and P95 time to first token of 4.26s across its model catalog. See the Models tab for per-model throughput and TTFT breakdowns.

Is DeepInfra reliable?

DeepInfra has a 50% success rate across 20 API calls in the last 7 days, with a 50% error rate.

Does DeepInfra support function calling?

Yes. 178 of 102 models on DeepInfra support function calling (tool use). The Capabilities tab lists which specific models accept tool definitions.

Does DeepInfra support JSON mode and structured output?

Yes. 167 of 102 models on DeepInfra support structured output (JSON mode / schema-constrained generation). The Capabilities tab shows which specific models accept response_format or json_schema parameters.

Does DeepInfra offer batch inference?

Yes. 189 of 102 models on DeepInfra support batch inference for cheaper, asynchronous workloads.

Does DeepInfra support multimodal models?

Yes. DeepInfra's catalog includes 53 vision-capable models. See the Models and Capabilities tabs for the full per-model breakdown.

Whose models does DeepInfra host?

DeepInfra hosts models from Anthropic, ByteDance, DeepSeek, Google, Gryphe, and IBM, plus 16 more. See the Models tab for the full catalog grouped by creator.

How do I start using DeepInfra?

Sign up at https://deepinfra.com/ to get an API key, then call DeepInfra's API directly from your application. Most clients work out of the box by pointing the OpenAI SDK at DeepInfra's base URL with your key. Use the Models and Pricing tabs above to pick the right model for your latency, cost, and context-window requirements.