DeepInfra API

LLM 5.7/10

Open models per-token: Llama 3.2 3B $0.02/M, Llama 8B $0.06/M, Qwen 2.5 72B $0.23/M, DeepSeek V3.1 $0.30/$1.00 per M tokens, gpt-oss-120b $0.09/$0.45 per M tokens, Qwen3-Coder $0.30/$1.20 per M tokens

DeepInfra API scores 5.7/10 for free value in LLM. The main constraint: Paid — Pay-as-you-go, from $0.02/M. Open models per-token: Llama 3.2 3B $0.02/M, Llama 8B $0.06/M, Qwen 2.5 72B $0.23/M, DeepSeek V3.1 $0.30/$1.00 per M tokens, gpt-oss-120b $0.09/$0.45 per M tokens, Qwen3-Coder $0.30/$1.20 per M tokens.

Category LLM Kind Paid plan

Free value score

5.7 / 10

Show the math

Model Quality 3.8/10

value = 21 artificial_analysis_score

weighted score = 0.2 * 3.8/10

Access 10/10

value = 10/10

weighted score = 0.1 * 10/10

Price Efficiency 5.6/10

value = 2105263 tok/$

weighted score = 0.7 * 5.6/10

Key metrics

DeepSeek V3.1 21 Best model

$0.475/M Price

What you get

Paid — Pay-as-you-go, from $0.02/M. Open models per-token: Llama 3.2 3B $0.02/M, Llama 8B $0.06/M, Qwen 2.5 72B $0.23/M, DeepSeek V3.1 $0.30/$1.00 per M tokens, gpt-oss-120b $0.09/$0.45 per M tokens, Qwen3-Coder $0.30/$1.20 per M tokens

Up to 50M tokens/$1 on small models; ~3.3M output/$1 on DeepSeek V3.1; ~11M input/$1 on gpt-oss-120b — broadest cheap catalog

Gotchas

No gotchas documented yet. Contribute one if you know a constraint we missed.

Sources

"deepseek-chat-v3.1 is listed at $0.30 input / $1.00 output per M tokens; gpt-oss-120b at $0.09 / $0.45 per M tokens; qwen3-coder at $0.30 / $1.20 per M tokens."
https://langdb.ai/app/providers/deepinfra
"At $0.23 per million tokens, Qwen 2.5 72B is roughly 1/10th the price of GPT-4o"
https://deepinfra.com/blog/qwen-api-pricing-2026-guide

paid value llm-providers

Open service ← All LLM

Alternatives

Same category · similar quota

Groq On-Demand APIGroq On-Demand API · paid · Paid plan

5.6

Llama 3.1 8B $0.05/$0.08; GPT-OSS $0.075/$0.30; Llama 3.3 70B $0.59/$0.79 per 1M. LPU ultra-fast inference

1.6M/$23.8 AA1.6M/$

May be outdatedChecked 26d ago

Paid onlyMetered

Command Code Provider APICommand Code Provider API · paid · Paid plan

5.4

Raw OpenAI- + Anthropic-compatible API ('no lock-in — any client or agent') with GLM-5.2 + GLM-5.2 Fast (1M ctx), Kimi K2.7 Code + HighSpeed, DeepSeek V4 Pro (75% off), MiMo V2.5 (98% off), plus Claude/GPT/Gemini at list; $15 usage included, then pay-as-you-go with permanent discounts

207.8K/$51.1 AA207.8K/$

May be outdatedChecked 23d ago

Paid onlyMetered

Moonshot Kimi APIMoonshot Kimi API · paid · Paid plan

6.0

Kimi K2/K2.5: $0.15/M input (Nov-2025 pricing, 75% off) to $0.60/M, $2.50-3.00/M output, 256K context; K2.5 adds multimodal

3.8M/$19.4 AA3.8M/$

May be outdatedChecked 26d ago

Paid onlyEmbeddingsMetered