GMI Cloud

LLM 5.4/10

GMI Cloud is an NVIDIA-partner inference cloud offering serverless open-model inference and H100/H200 GPUs, with free starter credits for new accounts.

Free credit: $5
Card required: No

GMI Cloud scores 5.4/10 for free value in LLM. The main constraint: Reported $5 in free credits for new users to try the Inference Engine; some free inference endpoints advertised without a card. Finite starter credit.

Category LLM Kind NVIDIA-powered inference + GPU cloud Caveats Free credit is finite (reported $5) and the offer wording shifts; sustained use is pay-as-you-go. Newer platform — verify the current free amount in-console. Card required No Eligibility anyone

Free value score

5.4 / 10

Show the math

Value 6.4/10

value = 5 $ one-time

weighted score = 0.7 * 6.4/10

Model Quality 1.9/10

value = 10 artificial_analysis_score

weighted score = 0.2 * 1.9/10

Volume 5.3/10

value = 14277556 tok/mo

weighted score = 0.1 * 5.3/10

Access 7/10

value = 7/10

score = 7/10

Key metrics

DeepSeek V3 10 Best model

$5 Credit

Yes Commercial use

Models

3 free · 230+ total

Limits may be stale — last re-verified outside this section's freshness window.

Model	Family	AA	Access
DeepSeek V3	DeepSeek	10.4	free
DeepSeek R1 Distill Qwen 1.5B	DeepSeek	—	free
Llama-3.1-8B-Instruct	Llama	—	free
Claude Opus 4.8	Claude	55.7	paid
GPT-5.5	GPT	54.8	paid
GPT-5.4	GPT	51.4	paid
Gemini 3.5 Flash	Gemini	50.2	paid
Gemini-3.1-Pro	Gemini	46.5	paid
DeepSeek-V4-Pro	DeepSeek	44.3	paid
GPT-5.2	GPT	42.2	paid
DeepSeek-V4-Flash	DeepSeek	40.3	paid
ZAI: GLM-5.1	GLM	40.2	paid
ZAI: GLM-5	GLM	39.5	paid
GPT-5.1	GPT	38.9	paid
MiniMaxAI/MiniMax-M2.7	MiniMax	38.1	paid
moonshotai/Kimi-K2.5	Kimi	38.1	paid
Claude Sonnet 4.6	Claude	35.9	paid
Claude Opus 4.5	Claude	34.7	paid
deepseek-ai/DeepSeek-V3.2	DeepSeek	24.7	paid
DeepSeek-V3.2-Exp	DeepSeek	24.7	paid
OpenAI: GPT OSS 120b	GPT	23.8	paid
GLM-4.6	GLM	23	paid
DeepSeek-V3.1-Terminus	DeepSeek	21.1	paid
DeepSeek R1	DeepSeek	20.1	paid
Kimi-K2-Thinking	Kimi	19.4	paid
GPT OSS 20b	GPT	14.9	paid
Llama-4 Maverick 17B 128E Instruct FP8	Llama	14.3	paid
DeepSeek V3 0324	DeepSeek	10.4	paid
Claude Haiku 4.5	Claude	—	paid
DeepSeek R1 Distill Llama 70B	DeepSeek	—	paid
deepseek-ai/DeepSeek-R1-0528	DeepSeek	—	paid
Llama 3.3 70B Instruct	Llama	—	paid
MiniMax-M2	MiniMax	—	paid
MiniMaxAI/MiniMax-M2.5	MiniMax	—	paid
moonshotai/Kimi-K2-Instruct-0905	Kimi	—	paid
Qwen: Qwen3-VL-235B-A22B-Instruct-FP8	Qwen	—	paid
Qwen3 235B A22B Instruct 2507 FP8	Qwen	—	paid
Qwen3 235B A22B Thinking 2507 FP8	Qwen	—	paid
Qwen3 Coder 480B A35B Instruct FP8	Qwen	—	paid
Qwen3 Next 80B A3B Thinking	Qwen	—	paid

Showing first 3 of a larger catalog.

Compare all free models across providers →

What you get

Reported $5 free credits for new accounts

NVIDIA H100/H200-backed serverless inference

Open-model catalog with OpenAI-compatible API

Also offers dedicated GPU clusters and bare metal

Gotchas

$5 starter credit (one-time)

New accounts receive a reported $5 in free credits for the Inference Engine; once exhausted the account switches to pay-as-you-go with no free allotment.
Card required after credit exhaustion

Continuing to use the platform after the $5 credit is depleted requires adding a payment method; a card may also be needed at signup to unlock full quota.

Sources

"By default, organizations are assigned to Tier 1. ... | Tier 1 | $0 | Immediately | ... | zai-org/GLM-5-FP8 | 100,000 | 2,000,000 | 4,000,000 | 10,000,000 | 150,000,000 |"
https://docs.gmicloud.ai/inference-engine/api-reference/rate-limit
"New users get $5 in free credits to try Inference Engine 2.0."
https://www.threads.com/@githubprojects/post/DQ4yj5ZkyOG/new-users-get-in-free-credits-to-try-inference-engine-join-the-gmi-community
"GLM-5 scores 50 on the Intelligence Index and is the new open weights leader, up from GLM-4.7's score of 42 - an 8 point jump"
https://x.com/ArtificialAnlys/status/2021678229418066004
"GMI Cloud, a San Jose-based startup that provides GPU cloud infrastructure, has raised an $82 million Series A led by Headline Asia."
https://techcrunch.com/2024/10/29/gmi-cloud-secures-82m-in-series-a-for-its-gpu-cloud-infrastructure/

gmi-cloud nvidia open-weight finite-credit gpu-cloud

Open service Source ← All LLM

Alternatives

Same category · similar quota

AI21 StudioAI21 Studio · finite_credit · First-party hosted LLM API

5.5

AI21 Studio gives new accounts a $10 trial credit (no card) good for three months to use the Jamba family of long-context hybrid SSM-Transformer models.

$10.0 one-time*5 AA2.9M tok

May be outdatedChecked 28d ago

One-time

b.aib.ai · finite_credit · LLM API aggregator

5.2

Third-party OpenAI-compatible aggregator with a reported large free-credit signup bonus.

$0.50 one-time*55.7 AA50K tok

May be outdatedChecked 28d ago

NoncomPersonalOne-time

Nebius AI StudioNebius AI Studio · trial · Hosted open-model LLM API

5.7

Nebius AI Studio offers serverless inference on a wide range of open models with a small free starter credit for new users, usable from playground or API after adding a bank card at signup.

$1.00 one-time*51.1 AA1.1M tok

May be outdatedChecked 28d ago

TrialOne-time