OVTH / 2026 LIVE · V0.4.0
Ø Overthinking Gateway ↗
OVTH Gateway · Nº 03

One endpoint.
18 models.
Zero config.

Dual-mode API: OpenAI-compatible chat completions and Anthropic-native messages behind one key. Use public model IDs; OVTH handles routing, retries, and compatibility at the gateway boundary.

quickstart · curl [copy]
# OpenAI-compatible
curl https://gateway.ovth.dev/v1/chat/completions \\
  -H "Authorization: Bearer ***" \\
  -H "Content-Type: application/json" \\
  -d '{
    "model": "gpt-5.5",
    "messages": [{"role":"user","content":"hi"}]
  }'
I.
Tiers · Four lanes, one key
001 / 005
Tier · Flagship
flagship/

Frontier reasoning + multimodal.

Models
5
Reasoning
5
Families
gpt-5.5 · claude-opus-4.7 · gemini-3.1-pro · grok-4.3 · gpt-5.3-codex
gpt-5.5
Tier · Mid
mid/

Daily-driver coding and research.

Models
6
Reasoning
6
Families
claude-sonnet-4.6 · deepseek-v4-pro · kimi-k2.6 · minimax-m2.7 · qwen3.5 · glm-5.1
claude-sonnet-4.6
Tier · Fast
fast/

Low-latency, low-cost throughput.

Models
5
Reasoning
5
Families
claude-haiku-4.5 · gpt-5.4-mini · deepseek-v4-flash · gemini-3-flash · mimo-v2-flash
deepseek-v4-flash
Tier · Open
open/

Community/open-weight lane.

Models
2
Reasoning
1
Families
gemma-4-31b-it · gpt-oss-120b
gpt-oss-120b
II.
Limiter · credit / 1M token cost
002 / 005
models
18
reasoning
17
native thinking
6
vision
11
1M+ context
5
model class context input / output · credits / 1M flags
gpt-5.5 flagship 1.05M 200 / 600 reasoning · vision
claude-opus-4.7 flagship 200K 600 / 3000 reasoning · vision
gemini-3.1-pro flagship 1.05M 140 / 420 reasoning · vision
grok-4.3 flagship 256K 120 / 600 reasoning
gpt-5.3-codex flagship 400K 200 / 600 reasoning · vision
claude-sonnet-4.6 mid 1M 120 / 600 reasoning · vision
deepseek-v4-pro mid 164K 60 / 200 reasoning
kimi-k2.6 mid 262K 60 / 200 reasoning · vision
minimax-m2.7 mid 1M 60 / 200 reasoning
qwen3.5-397b-a17b mid 262K 40 / 160 reasoning · vision
glm-5.1 mid 200K 40 / 160 reasoning
claude-haiku-4.5 fast 200K 40 / 200 reasoning · vision
gpt-5.4-mini fast 400K 16 / 64 reasoning · vision
deepseek-v4-flash fast 164K 10 / 40 reasoning
gemini-3-flash fast 1.05M 6 / 24 reasoning · vision
mimo-v2-flash fast 200K 6 / 24 reasoning
gemma-4-31b-it open 131K 0 / 0 vision
gpt-oss-120b open 131K 0 / 0 reasoning

Credit costs are shown per 1M input/output tokens. Capability flags reflect the public API surface; provider routing details are intentionally not exposed.

III.
Docs · OpenAI + Anthropic
003 / 005

Same SDK.
New superpowers.

Point OpenAI-compatible clients at /v1/chat/completions, or Claude/Anthropic clients at /v1/messages. Pick your model by public id; no provider keys or backend configuration are exposed to clients.

Full reference → /docs
GET /v1/models
All 18 canonical public chat models
GET /v1/limits
$/1M price + context + capabilities
POST /v1/chat/completions
OpenAI chat · streaming · vision · tool calls
POST /v1/messages
Anthropic messages · vision · thinking · tool_use
GET /health
Gateway status
GET /docs
Client reference · SDK examples
anthropic-native /v1/messages
# Anthropic-native Messages API
curl https://gateway.ovth.dev/v1/messages \\
  -H "Authorization: Bearer ***" \\
  -H "anthropic-version: 2023-06-01" \\
  -H "Content-Type: application/json" \\
  -d '{
    "model": "claude-sonnet-4.6",
    "max_tokens": 1024,
    "messages": [{"role":"user","content":"hi"}]
  }'
reasoning · max_tokens ≥ 512 tip
# reasoning / thinking
curl https://gateway.ovth.dev/v1/chat/completions \\
  -H "Authorization: Bearer ***" \\
  -H "Content-Type: application/json" \\
  -d '{
    "model": "claude-sonnet-4.6",
    "max_tokens": 1024,
    "reasoning_effort": "medium",
    "messages": [{"role":"user","content":"explain monads"}]
  }'
vision · image input url/base64
# vision · upload or public URL
curl https://gateway.ovth.dev/v1/chat/completions \
  -H "Authorization: Bearer *** \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [{
      "role": "user",
      "content": [
        {"type":"text","text":"describe this image"},
        {"type":"image_url","image_url":{"url":"https://example.com/image.png"}}
      ]
    }]
  }'

Uploads can be sent as base64/data URLs. Public image URLs work too when reachable from the gateway.

IV.
Pricing · Pact · Covenant · Oath
004 / 005
Plan · Trial
$0 /mo

Try it for 7 days.

Credits
3,000 credits · 7-day trial
Meter
Spend credits by model cost per 1M input/output tokens
Access
Full catalog · public ids only
Start free trial →
Plan · Pact
$4.99 /mo

For devs who explore.

Credits
3,000 credits / day · 90K / month
Meter
Spend credits by model cost per 1M input/output tokens
Access
Full catalog · best-effort
Subscribe →
Plan · Covenant
$9.99 /mo

For devs who ship.

Credits
6,000 credits / day · 180K / month
Meter
Spend credits by model cost per 1M input/output tokens
Access
Full catalog · priority queue
Subscribe →
Plan · Oath
$14.99 /mo

For devs who can't stop.

Credits
10,000 credits / day · 300K / month
Meter
Spend credits by model cost per 1M input/output tokens
Access
Full catalog · reserved · SLA
Subscribe →
Primary · QRIS
Auto self-serve

Pay directly via QRIS from the dashboard — works with any Indonesian bank or e-wallet. Plan activates in under a minute once payment confirms.

Alt methods · Contact
Bank transfer · Crypto · etc

Need another method or help? WhatsApp admin — usual response under an hour during work hours.

Legal
Monthly · no lock-in

Cancel anytime. Pro-rata refund on mid-cycle downgrades. Invoices and receipts auto-issued, stored on your account.