Kilo Code Gateway logo

Kilo Code Gateway

LLM 7.3/10 Jun 19, 2026

OpenAI-compatible API gateway from Kilo Code — one key to 300+ models, with a rotating free ":free" model tier and no credit card.

Requests / min
3.3
Card required
No

Kilo Code Gateway scores 7.3/10 for free value in LLM. The main constraint: Free ":free" models via the gateway: 200 requests/hour per IP, no card; OpenAI-compatible (base https://api.kilo.ai/api/gateway). Free roster rotates (~12 models, incl. auto-routers).

Category LLM Kind Coding model gateway Caveats ⚠ Free models route through partner trial endpoints (e.g. NVIDIA) that LOG and may TRAIN on your prompts/outputs — not for sensitive or production data. Free roster rotates and specific model names are volatile. Premium models require paid credits. Renewal daily Card required No Eligibility anyone

Free value score

7.3 / 10
Show the math
Model Quality 8.6/10
value = 48 artificial_analysis_score
weighted score = 0.5 * 8.6/10
Request Rate 2/10
value = 3.33 req/min
weighted score = 0.25 * 2/10
Access 10/10
value = 10/10
weighted score = 0.25 * 10/10

Key metrics

nvidia/nemotron-3-super-120b 48 Best model
3 R/min Requests/min
Unclear Commercial use

Models

3 free · 336 total

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

Model Family AA Access
kilo-auto/free Router — free
nvidia/nemotron-3-super-120b-a12b Nemotron — free
openrouter/free Router — free

Compare all free models across providers →

What you get

OpenAI-compatible gateway

Free tier, no credit card

200 requests/hour (free)

300+ models behind one key

Coding-optimized routing

Gotchas

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

Sources

  • "Kilo Code provides its own built-in API provider that gives you access to the latest frontier coding models... start immediately with free models, or add credits to access premium models."

    https://kilo.ai/docs/ai-providers/kilocode

  • "Free models are tagged with ":free" and priced at $0; anonymous access is allowed for free models, identified by IP. 200 requests/hour limit; HTTP 429 when exceeded."

    https://kilo.ai/docs/gateway/usage-and-billing

  • "OpenAI-compatible gateway — base URL https://api.kilo.ai/api/gateway; live model list at GET /api/gateway/models (filter isFree)."

    https://kilo.ai/docs/gateway
free-tier openai-compatible coding gateway
Open service Source ← All LLM

Alternatives

Same category · similar quota
Z.ai GLM Coding LiteZ.ai GLM Coding Lite · paid · Paid plan
7.3

GLM-5.2 + GLM-5-Turbo + GLM-4.7 in Claude Code/OpenCode/Cline via Anthropic- and OpenAI-compatible endpoints; ~80 prompts/5h, ~400 prompts/week (1 prompt = 15-20 model calls); 1 concurrent project

15×51.1 AA15.0×
May be outdatedChecked 23d ago
Best valuePaid only
Z.ai GLM Coding ProZ.ai GLM Coding Pro · paid · Paid plan
7.5

GLM-5.2 + GLM-5-Turbo; ~400 prompts/5h, ~2,000 prompts/week (5x Lite), faster generation + priority access; 1-2 concurrent projects; MCP 1,000 calls/mo

15×51.1 AA15.0×
May be outdatedChecked 23d ago
Best valuePaid only
Z.ai GLM Coding MaxZ.ai GLM Coding Max · paid · Paid plan
7.6

Highest GLM-5.2 quota: ~1,600 prompts/5h, ~8,000 prompts/week (20x Lite), dedicated resources during peak, first access to new models; 2+ concurrent projects; MCP 4,000 calls/mo

15×51.1 AA15.0×
May be outdatedChecked 23d ago
Best valuePaid only