Text and chat · OpenAI
GPT
OpenAI's models: Astra for the hardest work, Sol for coding and agents, Luna for cheap volume. Anonymized.
Anonymized High costWhat sets it apart
OpenAI’s line is split into three named tiers since GPT-5.6: Sol for demanding coding and agent work, Terra in the middle, Luna for cheap, fast, high-volume tasks at $0.10 per million input tokens. GPT-6 raises the window to 1.05M tokens with 128K output.
- Sol, Terra and Luna tiers
- 1.05M context, 128K output (GPT-6)
- Pro variants for harder problems
Recommended for: think through a hard problem or a long task.
Strengths and limits
Where it shines
- GPT-6.1 Sol: near Astra results at a quarter of its price
- Over 1M token context, vision and tool calling on GPT-6
- A tier for every budget, from Luna to Astra Pro
Think twice when
- Anonymized, not private: OpenAI's retention policy applies
- Pro versions of GPT-5.4 and 5.5 cost far more than anything else
- GPT-4o and GPT-5.2 are older: prefer GPT-6
- Ultrafast mode answers faster but costs about six times standard Sol
- GPT-6: 1.05M token context
- Vision, tool calling, reasoning effort
- "Pro" versions think longer on hard tasks
Use cases
Triaging a pile of tickets cheaply
Read these 200 support tickets and sort them into billing, login, bug and other. Give me a count per group.
GPT-6 Luna is priced for high-volume tasks, $0.10 per million input tokens.
An agent that edits a codebase
In this repo, replace the old date library with the standard one, run the tests and fix what breaks.
Sol is described by OpenAI as built for complex coding and agentic workflows.
Research on a hard question
Compare three ways to fund a small community library, with costs, risks and a recommendation for a town of 8,000 people.
Astra is the top tier for complex reasoning and research.
Against its rivals
- ClaudeClaude Opus 5.5 leads the index (58) over GPT-6 Astra (53), at less than half Astra’s price.#1 of 90 on Artificial Analysis Intelligence Index (58 points)
- GeminiGemini 4 Argon matches Astra at 53 on the index.#5 of 90 on Artificial Analysis Intelligence Index (53 points)
- GrokGrok 4.7 scores 46, below GPT-6.1 Sol at 52.#12 of 90 on Artificial Analysis Intelligence Index (46 points)
Version history
Each version, dated, with what it changed over the one before.
Hollow points are earlier versions no longer offered in Sub Rosa.
See the numbers
| Version | Released | points |
|---|---|---|
| GPT-4o mini | 2024-07 | 7 |
| GPT-4o | 2024-11 | 8 |
| GPT-5.2 | 2025-12 | 30 |
| GPT-5.3 Codex | 2026-02 | 33 |
| GPT-5.4 | 2026-03 | 39 |
| GPT-5.5 | 2026-04 | 38 |
| GPT-5.6 | 2026-07 | 47 |
| GPT-6 | 2026-09 | 53 |
| GPT-6.1 Sol | 2026-09 | 52 |
- GPT-4o mini
- Small, cheap model with 128K context at $0.15 per million input tokens.
- GPT-4o
- November 2024 snapshot of GPT-4o: text and image in, 128K context, non-reasoning.
- GPT-5.2
- Reasoning model with a 400K context window at $1.75 / $14 per million tokens.
- GPT-5.3 Codex
- Coding-specialized version, 400K context, scoring 33 on the Artificial Analysis index.
- GPT-5.4
- 1M token context window on the main model, and a mini tier with a 400K window at $0.75 per million input tokens.
- GPT-5.5
- Priced at $5 / $30 per million tokens; scores 38 on the index.
- GPT-5.6
- Splits the line into three named tiers: Sol ($4 / $20), Terra ($2 / $12) and Luna ($0.20 / $1.20), all with 1M context.
- GPT-6
- Adds Astra as the top tier ($10 / $50) alongside Sol and Luna, with a 1.05M window and 128K output.
- GPT-6 Luna costs $0.10 per million input tokens, half of GPT-5.6 Luna.
- GPT-6.1 SolLatest
- Sol refresh at $2 / $10 per million tokens, scoring 52 on the index against 48 for GPT-6 Sol.
- Also served in Ultrafast mode: the same model, answers up to 8 times faster at about six times the price.
Benchmarks
How to read these scores: what a benchmark tells you.
Artificial Analysis Intelligence Index: #4 of 90 models measured in the catalog. One score averaging about ten independent tests of reasoning, knowledge, math, coding and agent work, all run the same way for every model. Higher is better. Each model is listed at its highest reasoning effort. Two or three points apart is a close call; ten points is a different class of model.
See the numbers
| Model | points |
|---|---|
| 1. Claude Opus 5.5 | 58 |
| 2. Claude Sonnet 5.5 | 56 |
| 3. Claude Fable 5.1 | 53 |
| 4. GPT-6 Astra | 53 |
| 5. Gemini 4 Argon | 53 |
| 6. GPT-6.1 Sol | 52 |
| 7. Muse Wonder | 52 |
| 8. Claude Opus 5 | 51 |
| 9. Claude Fable 5 | 50 |
| 10. GPT-6 Sol | 48 |
All 21 models in this family
| Model | Use | Privacy | Price |
|---|---|---|---|
GPT-6.1 Solopenai-gpt-61-sol | Text | Anonymized | $4.496 per million tokens written$0.00315 a page written |
GPT-6 Solopenai-gpt-6-sol | Text | Anonymized | $4.496 per million tokens written$0.00315 a page written |
GPT-6 Astraopenai-gpt-6-astra | Text | Anonymized | $22.48 per million tokens written$0.0157 a page written |
GPT-6 Astra Proopenai-gpt-6-astra-pro | Text | Anonymized | $22.48 per million tokens written$0.0157 a page written |
GPT-6 Lunaopenai-gpt-6-luna | Text | Anonymized | $0.225 per million tokens written$0.000157 a page written |
GPT-5.6 Solopenai-gpt-56-sol | Text | Anonymized | $8.992 per million tokens written$0.00629 a page written |
GPT-5.6 Sol Proopenai-gpt-56-sol-pro | Text | Anonymized | $8.992 per million tokens written$0.00629 a page written |
GPT-5.6 Terraopenai-gpt-56-terra | Text | Anonymized | $5.395 per million tokens written$0.00378 a page written |
GPT-5.6 Terra Proopenai-gpt-56-terra-pro | Text | Anonymized | $5.395 per million tokens written$0.00378 a page written |
GPT-5.6 Lunaopenai-gpt-56-luna | Text | Anonymized | $0.54 per million tokens written$0.000378 a page written |
GPT-5.6 Luna Proopenai-gpt-56-luna-pro | Text | Anonymized | $0.54 per million tokens written$0.000378 a page written |
GPT-5.5openai-gpt-55 | Text | Anonymized | $13.49 per million tokens written$0.00944 a page written |
GPT-5.5 Proopenai-gpt-55-pro | Text | Anonymized | $80.93 per million tokens written$0.0567 a page written |
GPT-5.4openai-gpt-54 | Text | Anonymized | $6.744 per million tokens written$0.00472 a page written |
GPT-5.4 Proopenai-gpt-54-pro | Text | Anonymized | $80.93 per million tokens written$0.0567 a page written |
GPT-5.4 Miniopenai-gpt-54-mini | Text | Anonymized | $2.023 per million tokens written$0.00142 a page written |
GPT-5.3 Codexopenai-gpt-53-codex | Text | Anonymized | $6.295 per million tokens written$0.00441 a page written |
GPT-5.2openai-gpt-52 | Text | Anonymized | $6.295 per million tokens written$0.00441 a page written |
GPT-4oopenai-gpt-4o-2024-11-20 | Text | Anonymized | $4.496 per million tokens written$0.00315 a page written |
GPT-4o Miniopenai-gpt-4o-mini-2024-07-18 | Text | Anonymized | $0.27 per million tokens written$0.000189 a page written |
GPT-6.1 Sol Ultrafastopenai-gpt-61-sol-ultrafast | Text | Anonymized | $26.98 per million tokens written$0.0189 a page written |
Specifications
| Specification | GPT-6.1 Sol | GPT-6 Astra | GPT-6 Sol | GPT-6 Luna | GPT-5.6 Terra |
|---|---|---|---|---|---|
| Released | September 2026 | September 2026 | September 2026 | September 2026 | July 2026 |
| Weights | Closed | Closed | Closed | Closed | Closed |
| Accepts | text, images | text, images | text, images | text, images | text, images |
| Produces | text | text | text | text | text |
| Context | 1,050K tokens | 1,050K tokens | 1,050K tokens | 1,050K tokens | 1,000K tokens |
| Longest answer | 128K tokens | 128K tokens | 128K tokens | 128K tokens | · |
| Abilities | Reads images, Uses Sub Rosa’s tools, Thinks before answering | Reads images, Uses Sub Rosa’s tools, Thinks before answering | Reads images, Uses Sub Rosa’s tools, Thinks before answering | Reads images, Uses Sub Rosa’s tools, Thinks before answering | Reads images, Uses Sub Rosa’s tools, Thinks before answering |
| Price | $4.496 per million tokens written | $22.48 per million tokens written | $4.496 per million tokens written | $0.225 per million tokens written | $5.395 per million tokens written |
| Privacy | Anonymized | Anonymized | Anonymized | Anonymized | Anonymized |
Prices, context and lengths from the live catalog on October 16, 2026; parameters and licenses from the maker.
Sources
Everything on this page comes from these pages, read on October 16, 2026 or as dated on each score.
Show the 21 sources
- developers.openai.com/api/docs/models/gpt-6-astra
- developers.openai.com/api/docs/models/gpt-6-sol
- developers.openai.com/api/docs/models/gpt-6-luna
- artificialanalysis.ai/models/gpt-6-astra
- artificialanalysis.ai/models/gpt-6-1-sol
- artificialanalysis.ai/models/gpt-6-sol
- artificialanalysis.ai/models/gpt-6-luna
- artificialanalysis.ai/models/gpt-5-6-terra
- artificialanalysis.ai/models/gpt-5-6-sol
- artificialanalysis.ai/models/gpt-5-6-luna
- artificialanalysis.ai/models/gpt-5-5
- artificialanalysis.ai/models/gpt-5-4
- artificialanalysis.ai/models/gpt-5-4-mini
- artificialanalysis.ai/models/gpt-5-3-codex
- artificialanalysis.ai/models/gpt-5-2
- artificialanalysis.ai/models/gpt-4o
- artificialanalysis.ai/models/gpt-4o-mini
- artificialanalysis.ai/leaderboards/models
- datanorth.ai/news/openai-launches-gpt-6-sol-and-luna
- eesel.ai/blog/gpt-5-6
- developers.openai.com/api/docs/models/gpt-6.1-sol