GPT-6 Sol
OpenAI's September 2026 Sol model for coding and agentic work, since replaced by GPT-6.1 Sol
OpenAI has replaced this model with GPT-6.1 Sol.
Context window
1.05M
Input / 1M tokens
$2.00
Output / 1M tokens
$10.00
Provider
OpenAI
Cached input $0.20 and cache writes $2.50 per 1M (GPT-6.1 Sol lists cached input at $0.10). Max output 128K tokens. Knowledge cutoff April 20, 2026. · Data verified 2026-11-08
GPT-6 Sol is OpenAI's balanced model in the GPT-6 series, trained with the same methods as GPT-6 Astra. It brings Astra-level capabilities in factuality, coding, and professional tasks at significantly lower cost through improved inference and caching. Designed for demanding work including agentic coding, business automation, and software engineering. OpenAI's API model page now points to GPT-6.1 Sol as the newer Sol model, which has the same $2/$10 price with cheaper cached input ($0.10 vs $0.20 per 1M). GPT-6 Sol stays in the API and is still the model behind ChatGPT's Chat tab on paid plans.
Capability index
Relative estimates (0-100) to place this model against its peers, grounded in published benchmarks.
How to access it
Available via the OpenAI API (gpt-6-sol) on the Chat Completions, Responses and Batch endpoints, and on Amazon Bedrock (openai.gpt-6-sol). In ChatGPT it runs the Chat tab for Plus, Pro, Business and Enterprise users (GPT-6.1 Sol is offered only in ChatGPT Work and Codex).
Strengths
- ✓50% price reduction vs. GPT-5.6 Sol while maintaining comparable capability
- ✓Halved factual error rate compared to predecessor
- ✓Strong on coding and computer use tasks
- ✓Astra-level factual reliability at much lower cost
Best for developers who...
When to choose it (and when not to)
Reach for GPT-6 Sol when...
- →You want API behavior that matches the ChatGPT Chat model on paid plans
- →You have workloads tuned to GPT-6 Sol and want stable behavior
- →Running at scale with cost constraints
- →Balancing performance and affordability
Look elsewhere if...
- ✕New projects: GPT-6.1 Sol has the same price, cheaper cached input and stronger results
- ✕Requiring absolute maximum performance (use Astra instead)
- ✕Ultra-cost-sensitive work (use Luna instead)
How to use it
Quickstart
Pythonfrom openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-6-sol",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Use service_tier='fast' for priority processing at 2x the standard rate
API model id: gpt-6-sol
Benchmarks
| Benchmark | Score | Notes |
|---|---|---|
| AutomationBench | 33.2% | At maximum reasoning level, outperforms Claude Opus 5 at max effort at 9% of cost per task |
| Internal Factuality Evaluation | ~50% error reduction | Based on de-identified real ChatGPT conversations; half the errors of GPT-5.6 Sol |
Source: OpenAI Official Announcement
Compare GPT-6 Sol
Compare GPT-6 Sol with any other model
Build a comparison →All model comparisons →