DeepInfra

partially verified

trial credit · inference_host · confidence 70% · mixed · last verified 2026-08-17 · field guide →

Subject record

Base URLs	https://api.deepinfra.com/v1/openai openai-compatible
API key required	yes
Credit card	yes
Phone verification	unknown
May train on prompts	unknown
Quota	trial_credit [rpm, credits, concurrency] “Free-LLM quota table: '60 RPM \| $5 credit \| ~5M tokens (varies) \| One-time \| 40+ open-source models'. Official rate-limits docs: 'Default limit: 200 concurrent requests per model'; exceeding returns 'HTTP 429 with a Rate limited message'. Official pricing (re-fetched 2026-07-18): no free credit mentioned; 'You have to add a card or pre-pay or you won't be able to use our services.'”
Signup friction	steps unknown
Env var	DEEPINFRA_API_KEY
Notes	Pay-per-token host for open models ('you only pay for what you use. There are no long-term contracts or upfront costs' per official pricing). Categorized trial_credit on the strength of Free-LLM's $5-credit row, but official pages contradict the friction claim (card/pre-pay required) — treat as low-friction-cheap rather than truly free. OpenAI-compatible base URL https://api.deepinfra.com/v1/openai is DeepInfra's long-standing documented endpoint but was not re-confirmed on the two pages fetched this pass. cheahjs README does not list DeepInfra. Recheck 2026-07-18: $5-credit claim REFUTED on official pages — official pricing still documents no signup credit and mandates card/pre-pay; the claim survives only in third-party comparison sites. Category left trial_credit pending a live signup test, but routing policies should not assume free credit here.

Models (2)

model	capabilities	context	free	limits
deepseek-ai/DeepSeek-R1	chat, reasoning	?	trial	60 rpm
meta-llama/Meta-Llama-3.1-8B-Instruct	chat	?	trial	60 rpm

Provenance

source_deepinfra_docs	official_pricing	trust: high	checked 2026-07-18
source_repo_freellm	curated_repo	trust: low	checked 2026-07-18