Cerebras Inference

LLM 6.8/10 Jun 15, 2026

Free access to Cerebras-powered models with token-bucket rate limits.

Tokens / day: 1M
Requests / min: 5
Card required: No

Cerebras Inference scores 6.8/10 for free value in LLM. The main constraint: Free Trial examples: 5 RPM, 30K TPM, 1M TPH, 1M tokens/day (TPD) for listed models.

Category LLM Kind Fast inference Caveats Limits vary by model/tier and should be checked inside the account. Renewal monthly Card required No Eligibility anyone

Free value score

6.8 / 10

Show the math

Value 7.7/10

value = 22.1 $/mo

weighted score = 0.7 * 7.7/10

Model Quality 6.1/10

value = 34 artificial_analysis_score

weighted score = 0.2 * 6.1/10

Volume 6/10

value = 30000000 tok/mo

score = 6/10

Request Rate 2.4/10

value = 5 req/min

weighted score = 0.1 * 2.4/10

Access 10/10

value = 10/10

score = 10/10

Key metrics

Z.ai GLM 4.7 (zai-glm-4.7) 34 Best model

1.0M T/day Tokens/day

5 R/min Requests/min

Yes Commercial use

Models

2 free · 12 total

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

Model	Family	Context	Limits	AA	Access
Z.ai GLM 4.7 (zai-glm-4.7)	GLM	131.1K	30 rpm · 60K tpm · 1K rpd	33.8	free
OpenAI GPT OSS 120B (gpt-oss-120b)	GPT	131.1K	30 rpm · 60K tpm · 1K rpd	23.8	free
GPT-5.3-Codex-Spark	GPT	—	—	44.3	paid
Kimi K2.6	Kimi	—	—	42.8	paid
GLM 5.1	GLM	—	—	40.2	paid
GLM 5	GLM	—	—	39.5	paid
StepFun Step 3.7 Flash	Step	—	—	29.7	paid
Gemma 4 31B	Gemma	—	—	29.4	paid
MiniMax M2.5	MiniMax	—	—	—	paid
Qwen 3 235B A22B Instruct 2507 (qwen-3-235b-a22b-instruct-2507)	Qwen	—	—	—	paid
Qwen3 32B (qwen-3-32b)	Qwen	—	—	—	paid
StepFun Step 3.5 Flash	Step	—	—	—	paid

Compare all free models across providers →

What you get

Free Trial examples: 5 RPM

30K TPM / 1M TPH

1M tokens/day for listed models

Token-bucket limits

Gotchas

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

Sources

"| `gpt-oss-120b` | 5 | 30K | 1M | 1M | | `zai-glm-4.7` | 5 | 30K | 1M | 1M | (Free Trial: RPM / TPM / TPH / TPD)"
https://inference-docs.cerebras.ai/support/rate-limits
"Yes, Cerebras free tier provides 1 million tokens per day permanently with no credit card required. Sign up at cloud.cerebras.ai and start ..."
https://www.getaiperks.com/en/ai/cerebras-free-tier-guide
"GLM-4.7 (Reasoning) scores 42 on the Artificial Analysis Intelligence Index v4.0 ... most intelligent open weights model ... ahead of DeepSeek V3.2 (Reasoning, 41), Kimi K2 Thinking (40)"
https://x.com/ArtificialAnlys/status/2009117037667422457
"Fidelity and other investors valued Cerebras at $8.1 billion in a new funding round. The artificial intelligence chip company filed for an ... [IPO]; reported net income of $87.9 million on revenue of $510 million for 2025"
https://www.cnbc.com/2025/09/30/ai-chip-company-cerebras-raises-1-billion-in-pre-ipo-funding-round.html

free-trial fast open-models

Open service Source ← All LLM

Alternatives

Same category · similar quota

Cohere APICohere API · paid · Paid plan

6.7

Command-family models (Command R/R+, Command A) per-token for RAG/agents; trial key for testing, production keys metered per 1M tokens

6.7M/$29.3 AA6.7M/$

May be outdatedChecked 26d ago

EmbeddingsBest valuePaid only

NaraRouterNaraRouter · freemium · Multi-model LLM gateway

6.7

NaraRouter is a byNara AI gateway with an OpenAI-compatible API, Google registration, no credit card, and a free daily token allowance across three models.

$15.0/mo*33.6 AA150M/mo

Checked 1d ago

No card

Zhipu ZCodeZhipu ZCode · freemium · Coding agent + free model

7.0

Codex-like coding agent from Zhipu AI on its GLM models. Free access is a 5-day per-user trial (5M tokens/day, then throttled) — not a recurring monthly tier.

$18.4/mo*33.8 AA25M/mo

May be outdatedChecked 28d ago

TrialMay train