GPT-5.6 Luna Price Cut 80%: Benchmark Shows Superior Cost-Efficiency
OpenAI announced an 80% price cut for GPT-5.6 Luna, making it the most cost-efficient model in its class. Benchmarks show Luna surpasses Google's strongest model in intelligence index, costs less than mainstream open-weight models, and achieves over 60% faster inference at nearly 200 tok/s. This marks a shift from raw compute to cost-effectiveness, challenging open-source models' traditional cost advantage.
Confirmed
- Price cut and cost-efficiency: Official announcement confirms 80% reduction, making Luna the most price-efficient model. @BenBajarin's CS Bench tests show higher quality at lower cost than open-weight models. @Angaisb finds Luna costs about a quarter of Gemini 3.6 Flash on AAI total cost. @downingARK's chart shows Luna achieves similar intelligence score to Opus 5 low mode at pennies. @EverydayAI reports Luna matches Sonnet 5 in capability with 25x cost-effectiveness.
- Cost and speed drop: @charliermarsh notes token prices fell 13x in 4 months (GPT-5.5 $2.50/$15 vs Luna $0.20/$1.20). @AccBalanced adds actual serving cost below $3 even on B200 without optimization, though not fully passed to API endpoints. @brandongalang measures 60%+ speed increase to 200 tok/s.
- Code security: @cramforce cites Vercel AI Gateway's DeepsecBench showing Luna surpasses Sol in xhigh mode after price cut.
Unconfirmed
- @teortaxesTex claims Luna's cache read cost is 5.5x higher than DeepSeek V4-Pro in agentic scenarios, with long-sequence penalties and extra cache write fees, making DeepSeek more advantageous there. This is not yet verified by others.
Why it matters
- The drastic API price drop and dominant cost-efficient model lower AI adoption barriers. OpenAI's core research focuses on building highly efficient models for any intelligence level, expecting innovation at 'cheap to ignore' costs.
2026-07-30 ~ 2026-08-01 · 15 related posts
- Episode 1: GPT-5.6 Luna Price Cut 80%: Benchmark Shows Superior Cost-Efficiency(2026-07-30, 15 posts)
- Episode 2: Enabling Codex Max Reasoning Boosts Performance and Cuts Costs(2026-07-31, 3 posts)
- Episode 3: GPT-5.6 Benchmarks Leak: Sol Leads Performance, Luna Excels in Cost-Efficiency(2026-08-01, 3 posts)
- Episode 4: OpenAI Slashes GPT-5.6 Prices, Luna Down 80%(2026-08-01, 7 posts)
Primary sources
- LLM Inference Costs Drop Below $3 with B200s, Yet API Prices Stay High — AccBalanced · 2026-07-30
- [source] Benchmark: GPT 5.6 Luna Beats Open-Weight Models on Cost-Performance Ratio — BenBajarin · 2026-07-31
- LLM Inference Costs Plunge: Token Prices Drop to 1/13th in Four Months — charliermarsh · 2026-07-31
- GPT-5.6 Luna is Cheaper, Smarter, and Faster than Gemini 3.6 Flash — Angaisb_ · 2026-07-31
- GPT-5.6 Luna Outperforms Sol on Vercel Leaderboard After 80% Price Cut — cramforce · 2026-07-31
- Chart Shows LLMs Offer Incredible Intelligence Per Dollar — downingARK · 2026-07-31
- GPT-5.6 Luna Tested After 80% Price Drop: Blazing Fast Full-Stack Code Generation — BorisMPower · 2026-07-31
- Report: OpenAI's GPT-5.6 Luna Matches Sonnet 5 at 25x Cost Efficiency — EverydayAI_ · 2026-07-31
- Luna Beats DeepSeek in Performance, but Agentic Costs Remain High — teortaxesTex · 2026-07-31
- [source] Luna Model Prices Slashed by 80%, Ushering in Cheap Intelligence Era — teortaxesTex · 2026-07-31
- GPT 5.6 Luna Beats Google's Best in Intelligence and Undercuts Its Cheapest — Rare_Bunch4348 · 2026-07-31
- [source] Luna Model Hits Nearly 200 tok/s with 40% Price Drop, Outperforming 5.6 sol Fast Mode — brandon_galang · 2026-07-31
- OpenAI Offers March Flagship Intelligence at 1/13th the Price After 4 Months — _sholtodouglas · 2026-08-01
- GPT-5.6 Luna Price Drops 80%, Slashing Coding Task Costs by 60x — steipete · 2026-08-01
- GPT-5.6 Luna Price Cut by 80%, First-Party Synergy Boosts Performance — dfinke · 2026-08-01