Best AI LLM APIs & Models Tools

Anthropic's flagship model — stronger coding, agents, and honesty

Mistral's unified open-source model — reasoning + vision + coding, Apache 2.0

Mistral's 24B multimodal open-source model — beats GPT-4o Mini, Apache 2.0

Mistral's 24B latency-optimized open model — faster than Llama 3.3 70B, Apache 2.0

Mistral's 128B merged flagship — open weights, coding+reasoning+instructions

Mistral's MoE flagship + edge model family — Apache 2.0, multimodal, reasoning

Cohere's open-source agentic coding model — 30B MoE, 3B active, Apache 2.0

Mistral's low-latency code completion model — FIM, 80+ languages, 256k context

Mistral's code-specific embedding model — semantic code search and RAG for repos

Mistral's document OCR model — 74% win rate vs OCR 2, $2/1k pages, forms + handwriting

Mistral's SOTA open-weight coding model — 72.2% SWE-bench, free API

Mistral's first reasoning model — 73.6% AIME2024, native multilingual chain-of-thought, 10x faster via Le Chat

Enterprise GPT-4o-class performance at $0.40/$2.00 per million tokens — Mistral's mid-tier powerhouse

Mistral's 24B regional LLM — Arabic & South Asian languages, 150+ tok/s, self-hostable

Mistral × NVIDIA 12B open-weight model — 128k context, Tekken tokenizer, FP8 inference, Apache 2.0

Mistral's 7B Mamba-architecture coding model — linear-time inference, 256k context, Apache 2.0

Open-weight 7B math specialist from Mistral AI — STEM reasoning, MATH benchmark SOTA at release

Mistral's largest open-weights MoE — 141B total / 39B active, Apache 2.0

Anthropic's most capable model — state-of-the-art coding, vision, and long-horizon tasks

Captures traces from your production LLM work, then trains a cheaper open-weight model that beats the eval

Local-first least-cost inference router that keeps routine requests on your own hardware

One OpenAI-compatible API key for 75+ LLM, image, video, and audio models with smart routing and failover

Flat-rate, unmetered OpenAI-compatible API serving Qwen3.6-35B, priced by concurrency instead of tokens

Real-time sentiment API with confidence scores, emotions and a financial mode

One OpenAI-compatible key for GPT, Claude, Gemini, Sora and Flux, billed pay-as-you-go below list price

Fact-verification API returning confidence-scored verdicts with cited sources, for grounding LLM output

Credit-based gateway to image, video, text and music models from multiple vendors, from a $5 top-up

OpenAI-compatible gateway to 400+ models across Anthropic, OpenAI, Google, xAI and DeepSeek with spend analytics and per-key IAM, token pools from $30/month

One API for image, video and audio models with persistent artifact links and MCP support, free API key with up to 2,000 credits then usage-based

Daily-refreshed price index for 67 GPU providers and 26 LLM inference APIs with an agent-friendly JSON API, free key at 750 requests/day

Open-source RAG-as-a-service API built on Supabase row-level security — upload documents, query for cited context, self-host if you want

One credit-based API and MCP endpoint for Midjourney and 70+ other generative media models, with webhooks, a Studio UI and no per-seat fees

Conversational intelligence as an API — configurable extraction kits that return structured deal signals from call transcripts

Runs one prompt against 40+ models side by side with per-response cost and a routing recommendation, on one $15/mo seat

Prepaid GPT-5.x API, one key, from US$5

One OpenAI-compatible API for text, image, video, and voice models.

One API for leading AI models, with selected routes saving 90%+ versus provider reference prices.

Standard Compute — description pending review.

social media chat reconstruct — description pending review.