Releases·Tool Release
APIMart relays GPT-Image-2.5 and 300+ models through one OpenAI-compatible endpoint from $1 minimum
Two GitHub posts document APIMart, a relay gateway exposing 300+ image, video, and language models behind a single OpenAI-compatible endpoint at api.apimart.ai/v1. Listed pricing includes image2.5 (GPT-Image-2.5) from $0.0085 per 1K image, Seedance 2.5 from $0.0961/sec, and cached LLM input from $0.40/M, with pay-as-you-go billing from a $1 minimum. The posts also ship a reproducible framework comparing four gateway archetypes — self-hosted open source, managed control plane, relayed per-unit route, and direct vendor API — scored on protocol compatibility, retry/idempotency semantics, and observability. For practitioners, the pitch is swapping models by changing the 'model' field without client changes.
98 engagement2 sourcesWed, Sep 16, 2026, 07:13 AM
AI Overview
- APIMart exposes image2.5 (GPT-Image-2.5) at $0.0085 per 1K images, Seedance 2.5 at $0.0961/sec, and cached LLM input at $0.40/M via https://api.apimart.ai/v1.
- Gateway requires no subscription: users top up from $1, pay per call, and every response reports billed cost via cost/credits_cost fields.
- Gateway async pattern uses task_id submission with polling at GET /v1/tasks/{id}, plus retry, idempotency, timeout and failover conventions.
- One API key reaches 300+ image, video and language models; switching requires changing only the model field, not the client.
Related
Products·launched
AI API gateway/relay: 300+ models behind one OpenAI-compatible key with published per-call pricing
An API gateway/relay (apimart.ai) that exposes 307 models — LLMs, image generation, video generation — through a single OpenAI-compatible base URL and key, with published unit prices and per-call cost shown in every response. It targets developers who want pay-as-you-go USD access to many models without VPNs, multiple accounts, or minimum commitments beyond a $1 top-up.
An API gateway/relay (apimart.ai) that exposes 307 models — LLMs, image generation, video generation — through a single OpenAI-compatible base URL and key, with published unit prices and per-call cost shown in every response. It targets developers who want pay-as-you-go USD access to many models without VPNs, multiple accounts, or minimum commitments beyond a $1 top-up.
050309 engagement5 sourcesgithubThu, Sep 24, 2026, 05:05 PM
Releases·Tool Release
APIMart publishes AI API gateway pricing: Gemini 3 Pro $1.6 in/$9.6 out, GPT Image 2.5 $0.0085/image, Seedance 2.0 Mini $0.0106/sec
A set of multilingual GitHub repositories document APIMart, a third-party AI API gateway that fronts 307 models with a single OpenAI-compatible key. Published pricing (snapshot 2026-09-28) includes GPT Image 2.5 at $0.0085 per 1K image, Seedance 2.0 Mini at $0.01056 per 480P-second, and Qwen3.7 Flash input at $0.0228568/M tokens, with USD pay-as-you-go settlement and a $1 minimum top-up. The repos state the service has operated since last year with roughly 100,000 registered users, mostly enterprise accounts, and disclose no affiliation with any model vendor.
A set of multilingual GitHub repositories document APIMart, a third-party AI API gateway that fronts 307 models with a single OpenAI-compatible key. Published pricing (snapshot 2026-09-28) includes GPT Image 2.5 at $0.0085 per 1K image, Seedance 2.0 Mini at $0.01056 per 480P-second, and Qwen3.7 Flash input at $0.0228568/M tokens, with USD pay-as-you-go settlement and a $1 minimum top-up. The repos state the service has operated since last year with roughly 100,000 registered users, mostly enterprise accounts, and disclose no affiliation with any model vendor.
03201.7k engagement32 sourcesgithubMon, Sep 28, 2026, 12:00 AM
Social·Community
Claude Opus 4.7 API pricing: APIMart gateway lists $4 input, $0.40 cached input, $5 cache-write
A GitHub guide documents claude-opus-4.7 access through APIMart, a third-party OpenAI-compatible gateway, listing observed pricing of $4 input, $0.40 cached input, and $5 for 5-minute cache writes. The guide is not affiliated with the model vendor, so the pricing reflects one gateway's snapshot rather than official Anthropic rates.
A GitHub guide documents claude-opus-4.7 access through APIMart, a third-party OpenAI-compatible gateway, listing observed pricing of $4 input, $0.40 cached input, and $5 for 5-minute cache writes. The guide is not affiliated with the model vendor, so the pricing reflects one gateway's snapshot rather than official Anthropic rates.
56 engagement1 sourcegithubThu, Sep 24, 2026, 07:02 PM
Social·Community
GLM-5.3 Flash pricing listed on APIMart: $0.06 input, $0.012 cached input, $0.20 output per unit via OpenAI-compatible gateway
A GitHub guide documents how to call GLM-5.3-Flash through the APIMart third-party gateway, listing observed per-unit pricing and code examples. It is vendor-independent documentation for developers accessing the model over an OpenAI-compatible API.
A GitHub guide documents how to call GLM-5.3-Flash through the APIMart third-party gateway, listing observed per-unit pricing and code examples. It is vendor-independent documentation for developers accessing the model over an OpenAI-compatible API.
56 engagement1 sourcegithubThu, Sep 24, 2026, 06:57 PM
Products·launched
Community-curated directory of 310 AI API providers with free credits and a unified gateway
A community-curated directory of 310 AI API providers offering a combined $4.3k in free credits. Users can review, upvote, and edit provider listings to keep them accurate, and route any model through a single OpenAI/Claude-compatible gateway URL with failover and spend caps. It targets developers who want to claim free credits and access many LLM vendors without managing separate integrations.
A community-curated directory of 310 AI API providers offering a combined $4.3k in free credits. Users can review, upvote, and edit provider listings to keep them accurate, and route any model through a single OpenAI/Claude-compatible gateway URL with failover and spend caps. It targets developers who want to claim free credits and access many LLM vendors without managing separate integrations.
3 engagement1 sourcehackernewsThu, Sep 10, 2026, 10:00 AM