Quick answer
- Released
- September 3, 2026
- Superseded by
- GPT-6.1 Sol (September 29, 2026)
- API model
- gpt-6-sol
- Price
- $2 in / $10 out per MTok; $0.20 cached input
- Context
- 1,050,000 tokens (128K max output)
- Reasoning effort
- none to max
GPT-6 Sol: verified facts
| Field | Verified value |
|---|---|
| Vendor | OpenAI |
| Release | September 3, 2026 |
| API identifier | gpt-6-sol |
| Input / output price | $2 in / $10 out per MTok |
| Cache / long-context pricing | $0.20 cached input; long-context rates apply above 272K input tokens |
| Context / max output | 1,050,000 tokens (128K max output) |
| Knowledge cutoff | April 20, 2026 |
| Reasoning effort | none, low, medium, high, xhigh, max |
| Availability | Still selectable in Codex; OpenAI now points to GPT-6.1 Sol as the newer Sol model |
Every number on this page was checked against the linked vendor documentation on September 30, 2026. Prices are API list prices per million tokens; product subscriptions and regional availability are separate.
Your AI workspace for shipping software.
Download AIDEN free and point it at your existing Claude Code or Codex setup. No credit card, running in minutes.
Download AIDEN freeFree to start · macOS 12+ · No credit card required
What changed in GPT-6.1 Sol
| GPT-6 Sol | GPT-6.1 Sol | |
|---|---|---|
| Released | September 3, 2026 | September 29, 2026 |
| Input / output | $2 in / $10 out per MTok | $2 in / $10 out per MTok |
| Cached input | $0.20 (10% of input) | $0.10 (5% of input) |
| Reasoning effort | none to max | low to max |
| Knowledge cutoff | April 20, 2026 | April 30, 2026 |
| Tool calling | Responses API, or Chat Completions with effort none | Responses API |
The list price did not move, so the cached-input cut is the change a long agent session feels, because every turn resends a growing prompt prefix and that prefix is what caching discounts. Moving over has one trap: 6.1 Sol accepts neither none nor minimal, so a config pinned to none should move to low, and Chat Completions tool calls should move to the Responses API. The full model page is GPT-6.1 Sol for coding, and the Codex-side choice, including reasoning effort, is in the best Codex model guide.
Where GPT-6 Sol fits now
OpenAI built Sol for demanding reasoning and everyday complex coding. GPT-6 Sol shares Astra's 1.05M-token context and 128K maximum output at $2/$10 per million tokens against Astra's $10/$50. That everyday role now belongs to GPT-6.1 Sol, OpenAI's recommended Codex model. GPT-6 Sol is the one to keep in three cases:
Chat Completions tool calls
Workloads validated on it
Workspaces still waiting
| Strengths | Trade-offs |
|---|---|
| One fifth of Astra's base token price | Less capable than Astra on OpenAI's model-selection ladder |
| 1.05M-token context and 128K maximum output | Chat Completions function calling requires reasoning effort none |
| Supports reasoning from none through max in the Responses API | Requests above 272K input tokens use higher long-context rates |
Published benchmarks
| Benchmark | Score | Source / caveat |
|---|---|---|
| FrontierCode v1.1 (main) | 49.3% | Anthropic Sonnet 5.5 launch table, Opus 5.5: 54.4%; Sonnet 5.5: 52.1% at xhigh |