FULL STORY

GPT-5.6 Sol: From Channel Discounts to a Full Price Cut

After offering GPT-5.6 Sol at half price on OpenRouter and Vercel AI Gateway, OpenAI cut its API and credit prices by over 20% for three months, citing improved model capability and efficiency.

2026-08-17 ~ 2026-08-22 · 2 episodes · 11 posts

Episode 1 · OpenAI Cuts GPT-5.6 Sol Price 50% Exclusively on OpenRouter and Vercel (2026-08-17, 6 posts)

On August 17-18, OpenAI announced an exclusive 50% discount on GPT-5.6 Sol via OpenRouter and Vercel AI Gateway, bringing prices below its own official API. Following a similar move on the Luna model, the industry widely reads this as a market-share battle aimed at Anthropic.

Confirmed

  • Scope: OpenRouter and Vercel AI Gateway; the Vercel discount runs until September 18, 2026 (per @cramforce).
  • The discount applies to all token types, service tiers (including Fast mode) and regions (per @cramforce).
  • Discounted pricing: $2.50/1M input tokens, $15/1M output tokens (per @petrusenkomax).
  • The model scores above 90% on coding and reasoning benchmarks with 99.29% availability (per @petrusenkomax).

Analysis

  • SemiAnalysis (relayed by @zephyrz9 and @soumitrashukla9) calls it a clever marketing tactic: the two platforms account for a small share of OpenAI's token volume but are key sources for market-share estimates, so the cut can sway third-party stats.
  • @Hesamation notes a similar discount on Luna previously spiked usage; despite margin compression, the goal is winning developer API share against Anthropic.
  • @petrusenkomax adds that Amazon Bedrock leads on latency.

Why it matters

Targeted third-party discounts rather than a global price cut suggest OpenAI is cheaply influencing market-share metrics and developer mindshare; whether it is data-driven marketing or a routine promotion remains debated.

Episode 2 · OpenAI Cuts GPT-5.6 Sol Prices by Over 20% for Three Months (2026-08-22, 5 posts)

On August 22, OpenAI announced that API and credit prices for GPT-5.6 Sol will drop by more than 20% over the next three months, citing improved capabilities and efficiency. Some see it as a response to open-source model competition.