Verified model access

gpt-6-astra API

A prepaid Model.sale endpoint for gpt-6-astra: one API key, explicit limits and a live compatibility check. The requested model is never silently replaced.

Price$1.125 per 1M charged tokensInitial reservation $0.01, released after measured usage

Availability Live · healthyLast check 2026-10-02T23:29:20.632Z

Verified endpointsChat Completions · ResponsesStreaming is tested separately for each route.

Quick start

Create an API key, add balance and use the base URL below. Keep the key in an environment variable.

curl https://api.model.sale/v1/responses \
  -H "Authorization: Bearer $MODEL_SALE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6-astra","input":"Reply exactly OK"}'

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.model.sale/v1", api_key=os.environ["MODEL_SALE_API_KEY"])
response = client.responses.create(model="gpt-6-astra", input="Reply exactly OK")
print(response.output_text)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.model.sale/v1", apiKey: process.env.MODEL_SALE_API_KEY });
const response = await client.responses.create({ model: "gpt-6-astra", input: "Reply exactly OK" });
console.log(response.output_text);

Billing

The wallet reserves before dispatch and settles from terminal usage. Cached and reasoning token treatment follows the current billing policy.

Create API key

How to estimate a request

The public rate is a blended USD amount per one million charged tokens. It is not a separate input-token or output-token tariff. For a quick estimate, multiply the charged-token count by the rate and divide by 1,000,000. The request record is authoritative for actual usage and settlement.

Charged tokens	Illustrative charge
1,000	$0.001125
10,000	$0.01125
100,000	$0.1125

These are arithmetic examples, not a promise of a fixed per-request price. Actual token counts depend on your input, output and the model's reported usage. Any minimum request amount shown above is an initial reservation; after terminal usage is received, the wallet settles the measured charge and releases unused reserved funds.

LimitsYour API key controls RPM, TPM, concurrency and spend caps. See the dashboard before sending production traffic.

Errors401 usually means the key or authorization is invalid. A 429 indicates a rate or spend limit. For 5xx and timeouts, inspect Usage before retrying so the client does not duplicate work.

Need help?Use the setup guides and support form. Never include prompts, outputs or secrets in a ticket.

Before using this model in production

Confirm the exact model ID and one of the verified endpoints listed above.
Start with a dedicated API key and a low single-request and daily spend cap.
Send one short JSON request, then verify the completed request and charge in Usage.
If your client depends on streaming, test an SSE request and confirm that it reaches a terminal event.
Handle 401, 429, timeout and 5xx responses explicitly. A failed request is not automatically retried or switched to a different model by Model.sale.

Availability is a recent observation, not an uptime guarantee. Check the full model catalog, pricing table, client setup guides and billing methodology before choosing a production configuration.

ChatGPT API · Codex · GLM · DeepSeek

The full Model.sale catalog is loading its latest snapshot. ChatGPT-compatible API, Codex, GLM and DeepSeek use one prepaid key; live availability is checked separately so a refresh never silently changes your requested model.

Refreshing API capabilities, prices and model health…

ChatGPT API

Responses · Chat Completions · OpenAI SDK

CHECKING

Codex

CLI · App · VS Code · Responses

CHECKING

DeepSeek

Chat Completions · SSE streaming

CHECKING

GLM

Chat Completions · SSE streaming

CHECKING

Claude

Messages compatibility check

CHECKING

gpt-5.4gpt · compatibility check in progress

$0.10/ 1MCHECKING

gpt-5.4-20gpt · compatibility check in progress

$0.10/ 1MCHECKING

gpt-5.4-minigpt · compatibility check in progress

$0.09/ 1MCHECKING

gpt-5.5gpt · compatibility check in progress

$0.45/ 1MCHECKING

gpt-5.6-lunagpt · compatibility check in progress

$0.09/ 1MCHECKING

gpt-5.6-solgpt · compatibility check in progress

$0.45/ 1MCHECKING

gpt-5.6-terragpt · compatibility check in progress

$0.15/ 1MCHECKING

gpt-6-astragpt · compatibility check in progress

$1.125/ 1MCHECKING

gpt-6-lunagpt · compatibility check in progress

$0.08/ 1MCHECKING

gpt-6-solgpt · compatibility check in progress

$0.38/ 1MCHECKING

gpt-6.1-solgpt · compatibility check in progress

$0.40/ 1MCHECKING

claude-haiku-4-5claude · compatibility check in progress

$0.075/ 1MCHECKING

claude-opus-4-5claude · compatibility check in progress

$0.30/ 1MCHECKING

claude-opus-4-6claude · compatibility check in progress

$0.40/ 1MCHECKING

claude-opus-4-7claude · compatibility check in progress

$0.50/ 1MCHECKING

claude-opus-4-8claude · compatibility check in progress

$0.60/ 1MCHECKING

claude-sonnet-4-5claude · compatibility check in progress

$0.15/ 1MCHECKING

claude-sonnet-4-6claude · compatibility check in progress

$0.25/ 1MCHECKING

deepseek-flashdeepseek · compatibility check in progress

$0.03/ 1MCHECKING

deepseek-v4-flashdeepseek · compatibility check in progress

$0.20/ 1MCHECKING

deepseek-v4-prodeepseek · compatibility check in progress

$0.20/ 1MCHECKING

glm-5glm · compatibility check in progress

$0.12/ 1MCHECKING

glm-5.1glm · compatibility check in progress

$0.16/ 1MCHECKING

glm-5.2glm · compatibility check in progress

$0.20/ 1MCHECKING

glm-5.3glm · compatibility check in progress

$0.20/ 1MCHECKING

glm-5.3-flashglm · compatibility check in progress

$0.10/ 1MCHECKING

qwen-3.8qwen · compatibility check in progress

$0.25/ 1MCHECKING

qwen3.5-plusqwen · compatibility check in progress

$0.05/ 1MCHECKING

qwen3.6-plusqwen · compatibility check in progress

$0.04/ 1MCHECKING

qwen3.7-maxqwen · compatibility check in progress

$0.11/ 1MCHECKING

qwen3.7-plusqwen · compatibility check in progress

$0.03/ 1MCHECKING

kimi-k2.5kimi · compatibility check in progress

$0.07/ 1MCHECKING

kimi-k2.6kimi · compatibility check in progress

$0.11/ 1MCHECKING

kimi-k3kimi · compatibility check in progress

$0.24/ 1MCHECKING

mimo-v2-omnimimo · compatibility check in progress

$0.05/ 1MCHECKING

mimo-v2-promimo · compatibility check in progress

$0.05/ 1MCHECKING

mimo-v2.5mimo · compatibility check in progress

$0.01/ 1MCHECKING

mimo-v2.5-promimo · compatibility check in progress

$0.03/ 1MCHECKING

minimax-m2.5minimax · compatibility check in progress

$0.02/ 1MCHECKING

minimax-m2.7minimax · compatibility check in progress

$0.04/ 1MCHECKING

minimax-m3minimax · compatibility check in progress

$0.04/ 1MCHECKING

grok-4.6grok · compatibility check in progress

Price pending/ 1MCHECKING

gpt-4o-transcribemedia · compatibility check in progress

$0.10/ 1MCHECKING

gpt-image-1.5media · compatibility check in progress

$20.00/ 1MCHECKING

gpt-image-2media · compatibility check in progress

$0.10/ 1MCHECKING

gpt-image-2-codexmedia · compatibility check in progress

Price pending/ 1MCHECKING

gpt-image-2.5media · compatibility check in progress

$20.00/ 1MCHECKING

gpt-image-2.5-flaremedia · compatibility check in progress

$20.00/ 1MCHECKING

gpt-image-2.5-sunburstmedia · compatibility check in progress

$20.00/ 1MCHECKING

Prices and model IDs stay available during a refresh. Only a fresh successful check can make a model callable; an outage never silently substitutes another model.

Verified model access

gpt-6-astra API

A prepaid Model.sale endpoint for gpt-6-astra: one API key, explicit limits and a live compatibility check. The requested model is never silently replaced.

Price$1.125 per 1M charged tokensInitial reservation $0.01, released after measured usage

Availability Live · healthyLast check 2026-10-02T23:29:20.632Z

Verified endpointsChat Completions · ResponsesStreaming is tested separately for each route.

Quick start

Create an API key, add balance and use the base URL below. Keep the key in an environment variable.

curl https://api.model.sale/v1/responses \
  -H "Authorization: Bearer $MODEL_SALE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"gpt-6-astra","input":"Reply exactly OK"}'

import os
from openai import OpenAI

client = OpenAI(base_url="https://api.model.sale/v1", api_key=os.environ["MODEL_SALE_API_KEY"])
response = client.responses.create(model="gpt-6-astra", input="Reply exactly OK")
print(response.output_text)

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.model.sale/v1", apiKey: process.env.MODEL_SALE_API_KEY });
const response = await client.responses.create({ model: "gpt-6-astra", input: "Reply exactly OK" });
console.log(response.output_text);

Billing

The wallet reserves before dispatch and settles from terminal usage. Cached and reasoning token treatment follows the current billing policy.

Create API key

How to estimate a request

Charged tokens	Illustrative charge
1,000	$0.001125
10,000	$0.01125
100,000	$0.1125

LimitsYour API key controls RPM, TPM, concurrency and spend caps. See the dashboard before sending production traffic.

Errors401 usually means the key or authorization is invalid. A 429 indicates a rate or spend limit. For 5xx and timeouts, inspect Usage before retrying so the client does not duplicate work.

Need help?Use the setup guides and support form. Never include prompts, outputs or secrets in a ticket.

Before using this model in production

Confirm the exact model ID and one of the verified endpoints listed above.
Start with a dedicated API key and a low single-request and daily spend cap.
Send one short JSON request, then verify the completed request and charge in Usage.
If your client depends on streaming, test an SSE request and confirm that it reaches a terminal event.
Handle 401, 429, timeout and 5xx responses explicitly. A failed request is not automatically retried or switched to a different model by Model.sale.

Availability is a recent observation, not an uptime guarantee. Check the full model catalog, pricing table, client setup guides and billing methodology before choosing a production configuration.