Sam Altman urges US to lead in open-source and closed-source AI
Sam Altman told the US Senate that America must lead in both open-source and closed-source AI, with OpenAI planning an open model release this
TL;DR
OpenAI and Anthropic dramatically reduced prices for mid‑tier AI models, aiming to outcompete open‑source options and low‑cost Chinese rivals. The cuts could boost usage volumes but compress margins ahead of expected IPOs.
Key facts
- OpenAI cut GPT‑5.6 Luna price by 80% to $0.20 per million input tokens.
- OpenAI reduced GPT‑5.6 Terra price by 20% to $2 per million input tokens.
- Anthropic set permanent pricing for Claude Sonnet 5 at $2/$10 per million tokens.
- Anthropic introduced Claude Opus 5.5 at $4/$20 per million tokens with 40% effective savings.
- OpenAI launched GPT‑6 Sol and Luna at $2/$10 and $0.10/$0.50 per million tokens respectively.
- Chinese DeepSeek V4 Flash is priced at roughly $0.14/$0.28 per million tokens.
OpenAI and Anthropic announced aggressive price cuts on their mid‑tier AI models in a bid to outpace open‑source alternatives and Chinese competitors. On July 30, OpenAI reduced the cost of its GPT‑5.6 Luna model by 80%, setting the price at $0.20 per million input tokens and $1.20 per million output tokens, while also trimming its higher‑end GPT‑5.6 Terra model by 20% to $2 per million input and $12 per million output tokens. Anthropic followed on August 10 with permanent introductory pricing for Claude Sonnet 5 at $2/$10 per million tokens, and later introduced Claude Opus 5.5 at a list price of $4/$20 per million tokens. The moves come as both firms prepare for potential IPOs and face mounting pressure from low‑cost Chinese AI offerings.
Background
The AI pricing war entered a new phase in 2026 as OpenAI and Anthropic systematically lowered fees on their proprietary models. Historically, developers have balanced the convenience of hosted APIs against the operational overhead of self‑hosting open‑source models. Mid‑tier models, which handle routine tasks such as summarization and basic customer interactions, have become the battleground for price competition. OpenAI’s cuts on GPT‑5.6 Luna and Terra, followed by the launch of GPT‑6 Sol and Luna at even lower rates, aim to capture volume from enterprises that segment workloads by cost and performance. Anthropic’s pricing strategy mirrors this trend, cementing Claude Sonnet 5’s rates and introducing Claude Opus 5.5 with built‑in efficiency gains that translate to roughly 40% effective savings. Meanwhile, Chinese AI firms like DeepSeek are offering V4 Flash at roughly $0.14/$0.28 per million tokens, narrowing the performance gap and setting aggressive benchmarks that make even the reduced U.S. prices appear high. Both OpenAI and Anthropic are widely expected to pursue IPOs, betting that lower per‑call margins will be offset by higher usage volumes, a strategy that could reshape market dynamics for both proprietary and open‑source AI services.
What happened
On July 30, OpenAI announced an 80% price reduction for its GPT‑5.6 Luna model, bringing the cost down to $0.20 per million input tokens and $1.20 per million output tokens, and a 20% cut for GPT‑5.6 Terra to $2 per million input and $12 per million output tokens, as reported by CryptoBriefing. Anthropic then set permanent introductory pricing for Claude Sonnet 5 at $2 per million input tokens and $10 per million output tokens on August 10, according to the same source. On September 22, OpenAI launched GPT‑6 Sol and Luna with pricing of $2/$10 and $0.10/$0.50 per million tokens respectively, representing roughly a 50% reduction from previous generation pricing. Anthropic responded with Claude Opus 5.5 listed at $4 per million input tokens and $20 per million output tokens, while noting that efficiency improvements deliver around 40% effective savings for users. The article also highlighted that Chinese competitor DeepSeek V4 Flash is priced at roughly $0.14 per million input tokens and $0.28 per million output tokens, undercutting the U.S. offerings. Both OpenAI and Anthropic are expected to pursue IPOs, with the price cuts intended to boost adoption and usage metrics despite compressing margins.
Market impact
The price cuts introduced by OpenAI and Anthropic have reshaped cost expectations for AI API usage. OpenAI’s GPT‑5.6 Luna now costs $0.20 per million input tokens and $1.20 per million output tokens, while its GPT‑5.6 Terra is priced at $2/$12 per million tokens. Anthropic’s Claude Sonnet 5 is set at $2/$10 per million tokens, and Claude Opus 5.5 lists at $4/$20 per million tokens, though efficiency gains translate to about 40% lower effective costs for users. The September launch of OpenAI’s GPT‑6 Sol and Luna at $2/$10 and $0.10/$0.50 per million tokens respectively represents roughly a 50% reduction from prior generation pricing. In contrast, DeepSeek V4 Flash from China offers pricing around $0.14/$0.28 per million tokens, making even the reduced U.S. rates appear relatively expensive. The aggressive pricing is intended to drive higher volume usage, which could offset the thinner margins per API call and influence the valuation metrics for the anticipated IPOs of both firms.
Why it matters
The pricing war matters because it directly influences developer choices between proprietary hosted models and open‑source alternatives that require self‑hosting. By slashing fees, OpenAI and Anthropic aim to make their APIs more attractive for routine workloads, potentially pulling volume away from open‑source platforms like OpenRouter. The lower costs also intensify competition with Chinese AI firms that already offer ultra‑low pricing, forcing U.S. providers to reassess margin structures. For enterprises, the tiered approach—using cheap models for basic tasks and reserving premium models for specialized workloads—could become the new standard, reshaping spending patterns across the AI services market. Additionally, the price reductions are tied to upcoming IPO expectations, suggesting that both companies believe higher usage can compensate for reduced per‑call revenue, a strategy that could set a precedent for future AI market entrants.
DeFi Intel’s take
Editorial analysis, not a statement of fact.
DeFi Intel's take: the aggressive price cuts by OpenAI and Anthropic suggest a strategic shift toward volume‑driven growth as both firms eye IPOs. We read this as an effort to lock in enterprise usage before open‑source and Chinese competitors erode market share. The narrowing performance gap and the stark price differentials from Chinese models indicate that U.S. providers may need to further compress margins or innovate on efficiency to stay competitive. This environment could spur more hybrid AI stacks, where developers blend low‑cost proprietary APIs with open‑source models for niche tasks, potentially creating new arbitrage opportunities in the broader AI‑infrastructure ecosystem.
Sources used for this analysis
DeFi Intel wrote this analysis from the reporting above. Any figure or quotation can be checked against those pages.
Frequently asked questions
What price reductions did OpenAI implement for its GPT‑5.6 models?
OpenAI cut the price of GPT‑5.6 Luna by 80%, setting it at $0.20 per million input tokens and $1.20 per million output tokens, and reduced GPT‑5.6 Terra by 20% to $2 per million input tokens and $12 per million output tokens.
How does Anthropic's pricing for Claude Opus 5.5 compare to its earlier model?
Claude Opus 5.5 is listed at $4 per million input tokens and $20 per million output tokens, but built‑in efficiency improvements deliver around 40% effective savings for users compared with earlier pricing.
What pricing does the Chinese AI model DeepSeek V4 Flash offer?
DeepSeek V4 Flash is priced at roughly $0.14 per million input tokens and $0.28 per million output tokens, undercutting the reduced U.S. pricing from OpenAI and Anthropic.
Direct answers
Short answers, written to be quoted.
- What are the new prices for OpenAI's GPT‑5.6 Luna model?
- The new price for GPT‑5.6 Luna is $0.20 per million input tokens and $1.20 per million output tokens after an 80% cut.
- How does Anthropic's Claude Opus 5.5 pricing provide savings?
- Claude Opus 5.5 lists at $4 per million input tokens and $20 per million output tokens, but built‑in efficiency improvements deliver about 40% effective savings for users.
- Why are OpenAI and Anthropic cutting prices now?
- Both firms are cutting prices to boost adoption and usage metrics, aiming to increase volume to offset thinner margins as they prepare for potential IPOs.
DeFi Intel is an entity-graph aggregator: we curate, tag and link crypto news to a typed knowledge graph of protocols, tokens, people and incidents. We do not republish the full article body. Use the link above to read the original report at Crypto Briefing.
Entities in this story
Discussion
Comments are reviewed before publication.
Want the full article?
Continue reading on Crypto Briefing →
No comments yet.