OpenAI

GPT-6 Sol is OpenAI's September 22, 2026 model for complex coding and agentic workflows, positioned as a faster and cheaper GPT-6 lane than GPT-6 Astra. The API model ID is gpt-6-sol, with a 1,050,000-token context window, 128,000 max output tokens, text and image input, and reasoning.effort from none through max (default medium). Standard text pricing is $2 input and $10 output per million tokens.

Context tokens

1,050,000

Knowledge Cutoff

2026-04-20

Max output tokens

128,000

GPT-6 Luna is OpenAI's September 22, 2026 efficient model for focused, high-volume tasks in the GPT-6 family. The API model ID is gpt-6-luna, with a 1,050,000-token context window, 128,000 max output tokens, text and image input, and reasoning.effort from none through max (default medium). Standard text pricing is $0.10 input and $0.50 output per million tokens.

Context tokens

1,050,000

Knowledge Cutoff

2026-05-18

Max output tokens

128,000

GPT-6 Astra is OpenAI's September 2026 flagship for the hardest end-to-end reasoning, coding, computer use, research, and document work. The API model ID is gpt-6-astra, with a 1,050,000-token context window, 128,000 max output tokens, text and image input, and reasoning.effort from low through max. OpenAI's safety overview (September 3, 2026) states Astra is the first broadly deployed OpenAI model to reach the Critical cybersecurity capability level under the Preparedness Framework.

Context tokens

1,050,000

Knowledge Cutoff

2026-04-30

Max output tokens

128,000

OpenAI's GPT-5.6 Terra is the balanced GPT-5.6 tier for teams that want strong reasoning and coding quality at a lower cost than the Sol flagship lane.

GPT-5.6-Cyber is OpenAI's cybersecurity-specific model announced August 10, 2026 for Daybreak Red. Built on GPT-5.6 Sol, it is trained to improve specialized cyber tasks for trusted, authorized defenders—such as vulnerability research and exploit-chain development in approved environments—and to reduce refusals on certain higher-risk dual-use security prompts that general Sol still blocks.

Compact reasoning-focused model in OpenAI’s o-series line aimed at strong STEM and coding performance with lower cost than full o3. Intended for developers who want reasoning without always paying flagship prices—confirm exact API availability and snapshot names in OpenAI docs.

text-embedding-3-large produces high-dimensional text embeddings for semantic search, clustering, and classification. Teams pair it with pgvector or SaaS vector DBs for RAG; output dimensions can be reduced with tradeoffs described in OpenAI documentation.

Whisper large-v3 is OpenAI’s ASR model for transcription and translation across many languages, with strong robustness to accents and noise. It is commonly self-hosted or used via API partners; latency depends heavily on hardware and chunking strategy.

GPT-Image-2.5 Sunburst is OpenAI's September 8, 2026 most capable image generation and editing model. The API ID is gpt-image-2.5-sunburst, with dated snapshot gpt-image-2.5-sunburst-2026-09-08. It takes text and image input and outputs images. OpenAI positions it for workflows where editing precision matters most, with quality settings low, medium, high, xhigh, max, and auto. Call it on the Images API or as the model of the Responses API image generation tool.

API ID gpt-image-2.5-sunburst

GPT-Image-2.5 Flare is OpenAI's September 8, 2026 fastest GPT Image 2.5 model for high-quality everyday generation. The API ID is gpt-image-2.5-flare, with dated snapshot gpt-image-2.5-flare-2026-09-08. It accepts text and image input, outputs images, and supports the same quality ladder as Sunburst: low, medium, high, xhigh, max, and auto. Use it on the Images API or the Responses API image generation tool when latency matters more than Sunburst's editing-precision positioning.

API ID gpt-image-2.5-flare

GPT-Live-1 is OpenAI's full-duplex voice model for real-time spoken conversation. The API ID is gpt-live-1. It listens and speaks at the same time, handles interruptions, and delegates reasoning and tool use to a backend agent instead of doing that work itself. OpenAI's system card (July 8, 2026) introduced it as the paid ChatGPT voice default. The API launch post (September 10, 2026) documents Live sessions, native ASR transcripts and response text, keyword biasing, and turn detection. Voice sessions are billed $0.05 per minute, per second. Backend model and tool usage is billed separately.

Knowledge Cutoff

2025-07-31

Session Price Per Minute Usd

0.05

ChatGPT Images 2.5 is the September 8, 2026 ChatGPT image product for creating and editing images in conversation, the Images tab, Codex, and custom GPTs with Image Generation enabled. OpenAI's Images in ChatGPT help article says 2.5 brings sharper details, more precise editing, and faster generation, plus templates, mobile sketch-to-image, aspect-ratio control, and a selection editor. Images with thinking is available on Plus, Pro, and Business. The official DALL·E GPT was retired on August 30, 2026; ChatGPT Images replaces it.