OpenAI
GPT-6 Astra, GPT-5.6 Sol/Terra/Luna — plus image, voice, and realtime APIs
OpenAI's September 2026 API lineup spans GPT-6 Astra (flagship frontier), GPT-5.6 Sol (promotional), GPT-5.6 Terra (balanced), and GPT-5.6 Luna (everyday). Beyond text, the platform includes GPT-Image-2.5 (Sunburst and Flare), GPT-Live-1 voice, GPT-Realtime-2.1 for low-latency voice agents, and transcription/translation models. Batch API saves 50% on all models. Check the live table for current per-token rates.
A broad AI API platform — GPT-6 Astra to GPT-5.6 Luna covering multiple capability tiers, plus native image, voice, realtime, and transcription APIs under one key.
- Wide API surface: text, image, voice, realtime, transcription, translation, web search, code execution
- GPT-6 Astra at the frontier; GPT-5.6 Sol at promotional pricing
- Prompt caching on all major models (10% of input price)
- Batch API: 50% off all models for async workloads
- Wide ecosystem — virtually every AI framework supports OpenAI natively
- No open-weight models — full vendor lock-in
- GPT-6 Astra output pricing is among the highest for frontier tier — check live table
- Rate limits can be restrictive on lower tiers
OpenAI Models
| Model | Modality | Tier | Price | Cached /1M | TTFT | Throughput | Context | Intel. | Caps | |
|---|---|---|---|---|---|---|---|---|---|---|
dall-e-3 DALL-E | Image | advanced | $0.04 per image EST Jul 2026 | — | 10.0s | 0 t/s | 4K | 0 | ||
whisper-1 Whisper | Speech | efficient | $0.01 per minute EST Jul 2026 | — | 2.0s | 0 t/s | — | 0 | ||
text-embedding-3-small OpenAI Embeddings | Embedding | budget | $0.02/ $0.00 in / out per 1M EST Jul 2026 | — | 80ms | 0 t/s | 8K | 0 | ||
gpt-oss-20b GPT-OSS | Chat | budget | $0.04/ $0.15 in / out per 1M EST Jul 2026 | — | 200ms | 250 t/s | 131K | 62 | ||
gpt-5-nano GPT-5 | Chat | budget | $0.05/ $0.40 in / out per 1M EST Jul 2026 | — | 180ms | 280 t/s | 400K | 68 | ||
gpt-oss-120b GPT-OSS | Chat | efficient | $0.05/ $0.25 in / out per 1M EST Jul 2026 | — | 350ms | 160 t/s | 131K | 74 | ||
gpt-4.1-nano GPT-4.1 | Chat | budget | $0.10/ $0.40 in / out per 1M EST Jul 2026 | $0.03 | 180ms | 280 t/s | 1.0M | 48 | ||
text-embedding-ada-002 OpenAI Embeddings | Embedding | budget | $0.10/ $0.00 in / out per 1M EST Jul 2026 | — | 90ms | 0 t/s | 8K | 0 | ||
text-embedding-3-large OpenAI Embeddings | Embedding | efficient | $0.13/ $0.00 in / out per 1M EST Jul 2026 | — | 100ms | 0 t/s | 8K | 0 | ||
gpt-4o-mini GPT-4o | Chat | efficient | $0.15/ $0.60 in / out per 1M EST Jul 2026 | $0.07 | 280ms | 180 t/s | 128K | 58 | ||
gpt-5.4-nano GPT-5 | Chat | efficient | $0.20/ $1.25 in / out per 1M EST Oct 2026 | — | 200ms | 220 t/s | 1.0M | 78 | ||
gpt-5.6-luna GPT-5 | Chat | advanced | $0.20/ $1.20 in / out per 1M EST Aug 2026 | — | 350ms | 160 t/s | 1.1M | 88 | ||
gpt-5-mini GPT-5 | Chat | efficient | $0.25/ $2.00 in / out per 1M EST Jul 2026 | — | 300ms | 200 t/s | 400K | 80 | ||
gpt-4.1-mini GPT-4.1 | Chat | efficient | $0.40/ $1.60 in / out per 1M EST Jul 2026 | $0.10 | 300ms | 160 t/s | 1.0M | 62 | ||
gpt-3.5-turbo GPT-3.5 | Chat | budget | $0.50/ $1.50 in / out per 1M EST Jul 2026 | — | 200ms | 300 t/s | 16K | 35 | ||
gpt-4o-mini-realtime-preview GPT-4o Mini | Realtime | efficient | $0.60/ $2.40 in / out per 1M EST Jul 2026 | — | 250ms | 0 t/s | 128K | 0 | ||
gpt-5.4-mini GPT-5 | Chat | advanced | $0.75/ $4.50 in / out per 1M EST Oct 2026 | — | 350ms | 150 t/s | 1.0M | 85 | ||
o4-mini o4 | Chat | advanced | $1.10/ $4.40 in / out per 1M EST Jul 2026 | $0.28 | 2.5s | 95 t/s | 200K | 82 | ||
o3-mini o3 | Chat | advanced | $1.10/ $4.40 in / out per 1M EST Jul 2026 | $0.55 | 3.0s | 90 t/s | 200K | 80 | ||
gpt-5 GPT-5 | Chat | frontier | $1.25/ $10.00 in / out per 1M EST Jul 2026 | — | 600ms | 120 t/s | 400K | 92 | ||
gpt-5.2 GPT-5 | Chat | frontier | $1.75/ $14.00 in / out per 1M EST Jul 2026 | — | 550ms | 110 t/s | 400K | 90 | ||
o3 o3 | Chat | frontier | $2.00/ $8.00 in / out per 1M EST Jul 2026 | $0.50 | 8.0s | 45 t/s | 200K | 88 | ||
gpt-4.1 GPT-4.1 | Chat | frontier | $2.00/ $8.00 in / out per 1M EST Jul 2026 | $0.50 | 600ms | 100 t/s | 1.0M | 80 | ||
gpt-5.6-terra GPT-5 | Chat | frontier | $2.00/ $12.00 in / out per 1M EST Aug 2026 | — | 650ms | 110 t/s | 1.1M | 93 | ||
o4-mini-deep-research o4 | Chat | advanced | $2.00/ $8.00 in / out per 1M EST Jul 2026 | — | 30.0s | 40 t/s | 200K | 83 | ||
gpt-4o GPT-4o | Chat | frontier | $2.50/ $10.00 in / out per 1M EST Jul 2026 | $1.25 | 520ms | 110 t/s | 128K | 78 | ||
gpt-5.4 GPT-5 | Chat | frontier | $2.50/ $15.00 in / out per 1M EST Jul 2026 | — | 700ms | 100 t/s | 1.0M | 93 | ||
gpt-4o-transcribe GPT-4o | Speech | frontier | $2.50/ $10.00 in / out per 1M EST Jul 2026 | — | 1.5s | 0 t/s | — | 0 | ||
gpt-5.6-sol GPT-5 | Chat | frontier | $4.00/ $20.00 in / out per 1M EST Aug 2026 | — | 900ms | 70 t/s | 1.1M | 95 | ||
gpt-6.1-sol GPT-6 | Chat | frontier | $4.00/ $20.00 in / out per 1M EST Sep 2026 | $2.00 | 700ms | 100 t/s | 1.2M | 95 | ||
gpt-5.5 GPT-5 | Chat | frontier | $5.00/ $30.00 in / out per 1M EST Jul 2026 | — | 800ms | 90 t/s | 1.0M | 94 | ||
gpt-image-1 OpenAI Image | Image | frontier | $5.00/ $40.00 in / out per 1M EST Jul 2026 | — | 8.0s | 0 t/s | 33K | 0 | ||
gpt-4o-realtime-preview GPT-4o | Realtime | frontier | $5.00/ $20.00 in / out per 1M EST Jul 2026 | — | 300ms | 0 t/s | 128K | 0 | ||
gpt-4-turbo GPT-4 | Chat | advanced | $10.00/ $30.00 in / out per 1M EST Jul 2026 | — | 700ms | 80 t/s | 128K | 72 | ||
o3-deep-research o3 | Chat | frontier | $10.00/ $40.00 in / out per 1M EST Jul 2026 | — | 60.0s | 15 t/s | 200K | 90 | ||
gpt-6 GPT-6 | Chat | frontier | $10.00/ $50.00 in / out per 1M EST Sep 2026 | $5.00 | 900ms | 80 t/s | 1.2M | 98 | ||
o1 o1 | Chat | frontier | $15.00/ $60.00 in / out per 1M EST Jul 2026 | $7.50 | 12.0s | 35 t/s | 200K | 86 | ||
gpt-5-pro GPT-5 | Chat | frontier | $15.00/ $120.00 in / out per 1M EST Oct 2026 | — | 1.0s | 65 t/s | 400K | 94 | ||
o3-pro o3 | Chat | frontier | $20.00/ $80.00 in / out per 1M EST Jul 2026 | — | 20.0s | 30 t/s | 200K | 92 | ||
gpt-5.2-pro GPT-5 | Chat | frontier | $21.00/ $168.00 in / out per 1M EST Oct 2026 | — | 1.1s | 60 t/s | 1.0M | 95 | ||
gpt-5.4-pro GPT-5 | Chat | frontier | $30.00/ $180.00 in / out per 1M EST Jul 2026 | — | 1.2s | 60 t/s | 1.0M | 96 | ||
gpt-5.5-pro GPT-5 | Chat | frontier | $30.00/ $180.00 in / out per 1M EST Oct 2026 | — | 1.2s | 55 t/s | 1.0M | 96 | ||
gpt-4 GPT-4 | Chat | advanced | $30.00/ $60.00 in / out per 1M EST Jul 2026 | — | 900ms | 50 t/s | 8K | 65 | ||
o1-pro o1 | Chat | frontier | $150.00/ $600.00 in / out per 1M EST Jul 2026 | — | 30.0s | 20 t/s | 200K | 91 |
OpenAI — Frequently Asked Questions (FAQs)
Similar Providers
OpenAI Reviews
All LLM Providers
Full comparison tableOpenAI LLM pricing overview
OpenAI is a proprietary AI lab that develops and hosts its own large language models. OpenAI's September 2026 API lineup spans GPT-6 Astra (flagship frontier), GPT-5.6 Sol (promotional), GPT-5.6 Terra (balanced), and GPT-5.6 Luna (everyday). Beyond text, the platform includes GPT-Image-2.5 (Sunburst and Flare), GPT-Live-1 voice, GPT-Realtime-2.1 for low-latency voice agents, and transcription/translation models. Batch API saves 50% on all models. Check the live table for current per-token rates. Common use cases include Coding, Chat, Vision, Reasoning. Headquartered in San Francisco, CA, founded 2015. All models are accessible via a REST API compatible with standard OpenAI-style request formats, enabling drop-in integration with most LLM frameworks and orchestration tools.
OpenAI vs other LLM providers
OpenAI competes with OpenAI, Anthropic, Google, Groq, Together AI, Mistral, Cohere, and other inference API providers across dimensions of price, throughput, latency, context window, and model intelligence. A broad AI API platform — GPT-6 Astra to GPT-5.6 Luna covering multiple capability tiers, plus native image, voice, realtime, and transcription APIs under one key. Use the LLM provider comparison tool to see OpenAI token pricing, latency, and throughput side-by-side with any other provider. The full LLM pricing table shows all providers ranked by input token cost, output token cost, and throughput in a single sortable view.
Why choose OpenAI?
OpenAI stands out because: A broad AI API platform — GPT-6 Astra to GPT-5.6 Luna covering multiple capability tiers, plus native image, voice, realtime, and transcription APIs under one key.. The API currently offers 44 models, spanning 22 frontier, 8 advanced, 8 efficient, 6 budget tiers. For teams running high-volume inference workloads, prompt caching and batch API endpoints can reduce effective input token costs by 50–90% — check the context window cost guide for a full breakdown of caching economics.
Understanding OpenAI token pricing
OpenAI charges separately for input (prompt) and output (completion) tokens, priced per 1M tokens in USD. Output tokens are typically 3–5× more expensive than input tokens due to the compute cost of autoregressive generation. The cheapest {meta.name} input rate is currently $0.04/1M tokens. Prices shown are sourced from OpenAI's public pricing page and updated every 15 minutes. Need help estimating your spend? Read the LLM API cost calculator guide — it covers tokens, context windows, prompt caching, and batch discounts with worked examples for RAG, chat history, and document processing workloads.
OpenAI context window and model capabilities
Context window size directly affects both capability and cost — every token in the context window is charged as an input token. OpenAI's largest context window is 1.2M tokens. For RAG and document processing workloads, longer context windows enable richer retrieval but increase per-call costs proportionally. Prompt caching — where supported — stores the KV state of repeated prefixes and charges 75–90% less for cache hits, making it the most impactful cost optimization for applications with consistent system prompts or retrieved documents. See the LLM context window cost guide for a full analysis of how context length affects your API bill.
Self-hosted vs managed inference: when OpenAI makes sense
Managed inference APIs like OpenAI eliminate infrastructure overhead — no GPU provisioning, driver management, or model serving stack to maintain. The trade-off is cost at scale: a single H100 at ~$2.50/hr can serve ~500K tokens/min of Llama 3.3 70B, which at OpenAI API rates would cost significantly more per token. The break-even point depends on your request volume, latency requirements, and engineering capacity. For teams processing fewer than ~10M tokens/day, managed APIs are almost always cheaper when total cost of ownership is considered. Above that threshold, self-hosted inference on rented GPU compute typically wins on unit economics. Read the cheapest GPU cloud guide for a full break-even analysis. Historical OpenAI token price data is available in the LLM price history charts.
Compare OpenAI With Another LLM Provider
Compare OpenAI vs other providers
Side-by-side token pricing, latency, throughput, and context window. View all LLM provider comparisons →
Looking for alternatives?
Compare OpenAI with other LLM API providers side by side.