GPT-5.6 Sol Now $4/$20 per 1M Tokens, 33% Cheaper Output
OpenAI's GPT-5.6 Sol now lists at $4 per 1M input and $20 per 1M output tokens, undercutting Anthropic's Claude Opus 5 on output pricing. The API price reduction follows cuts to Terra and Luna as AI labs compete on inference economics.
Beat this week
Last 7 days Ā· AI Models
Impact 6.0/10 (+0.2 vs prior). Counts are stories in our record, not a market forecast.
Open the change reportCoverage balance Positive coverage leads. Positive coverage exceeds negative coverage by 46 percentage points.
This story sits in AI Models ā the counts compare this beat's last 7 days with the previous 7 in our verified record, not a market forecast.
Figures are computed live from our source-verified story record (as of ) The volume change compares this window with the prior 7 days in the same record. ā see our methodology for how impact and sentiment are derived.
AI briefing
Key takeaways
- OpenAI's GPT-5.6 Sol now lists at $4 per 1M input and $20 per 1M output tokens, undercutting Anthropic's Claude Opus 5 on output pricing.
- The API price reduction follows cuts to Terra and Luna as AI labs compete on inference economics.
- texasguardian.com
- floridastatesman.com
- irishsun.com
In this briefing
Mentioned
Key Intelligence
Key Facts
- 1OpenAI cut GPT-5.6 Sol API pricing to $4 per 1 million input tokens and $20 per 1 million output tokens for standard short-context use, down from $5 and $30 respectively.
- 2The reduction equals a 20 percent decrease on input tokens and a 33 percent decrease on output tokens.
- 3New pricing applies to the OpenAI API and eligible credits on ChatGPT Work and Codex, while Pro, Plus, and Business subscription prices remain unchanged.
- 4Late last month, OpenAI cut GPT-5.6 Terra pricing by 20 percent and Luna pricing by 80 percent.
- 5Anthropic's Claude Opus 5 lists at $5 per 1 million input tokens and $25 per 1 million output tokens; Claude Fable 5 lists at $10 and $50.
- 6OpenAI announced the pricing change on August 24, 2026 according to the company's pricing table.
| Model | ||
|---|---|---|
| GPT-5.6 Sol (new) | $4 | $20 |
| Claude Opus 5 | $5 | $25 |
| Claude Fable 5 | $10 | $50 |
| GPT-5.6 Sol (old) | $5 | $30 |
Analysis
AI engineers comparing frontier model APIs now have a new cost-performance baseline: GPT-5.6 Sol at $4 per million input tokens and $20 per million output tokens. The 33 percent output-cost reduction reshapes model selection for agentic workflows, coding agents, and retrieval-heavy systems, where output expense dominates.
OpenAI moved on August 24, 2026 to cut developer pricing for its GPT-5.6 Sol model, reducing standard short-context API rates from $5 to $4 per 1 million input tokens and from $30 to $20 per 1 million output tokens. That is a 20 percent cut on inputs and a 33 percent cut on outputs, and it applies immediately to the OpenAI API, with eligible credits rolling out across ChatGPT Work, the company's agentic AI product, and Codex, its coding tool. Consumer-facing Pro, Plus, and Business subscription prices were left unchanged, signaling that this is a developer- and platform-focused price move rather than a broad consumer discount.
The Sol reduction now creates a clear cost gap with Anthropic: Claude Opus 5 is listed at $5 per 1 million input tokens and $25 per 1 million output tokens, while frontier-tier Claude Fable 5 is listed at $10 and $50.
The announcement is part of a broader repricing wave across OpenAI's GPT-5.6 family. Late last month, the company reduced prices for its mid-tier GPT-5.6 Terra model by 20 percent and for its lower-cost Luna model by 80 percent. The Sol reduction now creates a clear cost gap with Anthropic: Claude Opus 5 is listed at $5 per 1 million input tokens and $25 per 1 million output tokens, while frontier-tier Claude Fable 5 is listed at $10 and $50. For developers selecting between frontier-class APIs, OpenAI's new list pricing is now the cheapest on output, which is often the dominant cost in generative tasks, and close to the lowest on input.
The immediate beneficiary is the AI-native SaaS ecosystem. For SaaS platforms that embed GPT-5.6 Sol in coding assistants, agentic automation, document processing, or customer support, output tokens typically represent the largest variable infrastructure expense. A one-third reduction in output price can materially expand gross margins or make previously uneconomic features viable at current per-seat pricing. The inclusion of ChatGPT Work and Codex credits means that existing enterprise software buyers may see an additional effective cost reduction, though OpenAI did not detail the exact credit mechanics. In practical terms, a customer-support automation handling 1 billion output tokens per month would see its Sol output cost fall from about $30,000 to $20,000, excluding any volume discounts or credits.
What to Watch
OpenAI's strategy is classic platform pricing. By lowering unit prices while the industry is still expanding, it is prioritizing developer adoption, usage volume, and ecosystem lock-in over near-term API revenue per token. Historically, AI and cloud inference has shown high usage elasticity: as prices fall, developers send more tokens, add longer context windows, and run more agentic loops, which can offset lower unit prices and grow total revenue. The move also puts pressure on Anthropic and other frontier competitors to respond with their own price cuts or differentiation claims. If OpenAI can pair lower list prices with quality and latency parity, it may push Claude Fable 5 and Claude Opus 5 into a difficult pricing corner.
Looking ahead, the developer pricing environment is likely to continue deflating as model efficiency improves and competition intensifies. OpenAI may follow with reduced pricing for fine-tuning, batch processing, and longer-context variants, while Anthropic and Google may respond with matching cuts, committed-use discounts, or free-tier expansions. Such responses would reinforce cheaper inference as a default assumption for planning. For SaaS operators, token costs should be modeled as a declining input rather than a fixed cost line, which changes build-versus-buy, context-length, and agent-depth decisions. For AI investors and strategists, frontier API list prices falling this quickly are a reminder that model-level margins are not guaranteed, and the durable value may flow to application layers that capture workflow and distribution. The bottom line is that lower token prices are good for developers but heighten the strategic pressure on model providers to monetize through premium enterprise features, agents, and workflow ownership rather than raw token arbitrage. Contract negotiations should account for this repricing tempo.
Source cluster
Primary reporting
- texasguardian.comOpenAI lowers developer pricing for GPT - 5 . 6 Sol model
- floridastatesman.comOpenAI lowers developer pricing for GPT - 5 . 6 Sol model
Cite This Page
"GPT-5.6 Sol Now $4/$20 per 1M Tokens, 33% Cheaper Output." AI Intelligence Brief, August 24, 2026. https://getaibrief.com/story/openai-gpt-5-6-sol-price-drop-vs-anthropic
How we covered this story
Every story in our AI coverage is assembled from multiple primary sources, cross-referenced for factual consistency, and scored along three independent dimensions: sentiment, operational impact, and source-cluster confidence. Single-source rumors and unverifiable claims do not pass our editorial gate. When a story shows "Verified by N sources" with Nā„2, the development is independently corroborated; when N=1, we mark it explicitly so readers can weigh the signal accordingly.
Impact scoring uses a 1-10 scale weighted toward regulatory, financial, and operational consequence rather than coverage volume. A topic that runs in every outlet but moves no real decisions ranks lower than a niche regulatory filing that reshapes how operators in the AI space have to behave. Read our full methodology for the scoring rubric, our glossary for term definitions, and our trends index for the longitudinal view across the beat.
Sources are only linked to a story once they clear our classification pipeline at a minimum 35 percent relevance threshold. According to that methodology, reviewed July 2026, this follows multi-source corroboration standards recommended by journalism research bodies such as the Reuters Institute for the Study of Journalism.
See something wrong in this story ā a wrong fact, a broken source link, a misattributed entity? Report a data issue.
| Signal on this page | What it tells you |
|---|---|
| Verified by N sources | Independent corroboration count. Nā„2 is our confidence floor; N=1 is marked explicitly. |
| Impact score (1-10) | Regulatory + financial + operational weight. 8+ signals an experienced-operator action item. |
| Sentiment | Five-tier classification trained on labeled AI-specific corpora. |
| Timeline | Where applicable, the related-events sequence that contextualizes today's development. |