OpenAI Launches GPT-6 Sol and Luna at Half Price; Big Hallucination Gains but Not a Clean Sweep
On September 23, OpenAI officially launched GPT-6 Sol and GPT-6 Luna, available the same day in ChatGPT Work and Codex (covering Plus/Pro/Business/Enterprise/Edu, with Free and Go users getting Luna on desktop) plus API access (gpt-6-sol). Sam Altman claimed across-the-board gains in intelligence, alignment, coding and computer use at half the per-token price of GPT-5.6. Artificial Analysis benchmarks and community tests, however, show the new models are not a clean sweep: cheaper, more efficient and far less hallucination-prone, but regressing on some software-engineering and knowledge-work benchmarks. Rival releases from Anthropic and Qwen the same day sharpened the coding-API price war, which became a community meme (ns123abc's "Sir" template).
Confirmed
- Halved pricing: Sol dropped from $4/$20 to $2/$10 per million tokens, Luna from $0.20/$1.20 to $0.10/$0.50. OpenAI claims some coding-eval task costs are up to 93% cheaper than Claude Opus 5 (an official figure relayed by Polymarket, pending third-party verification); Altman called per-task pricing unbeaten in the market. The official positioning is "build with Sol, scale with Luna," both models building on GPT-6 Astra with stronger coding/computer use, higher factuality and clearer, jargon-light answers; Sol targets Claude Opus 5.5 while Luna emphasizes cost efficiency.
- Coding agent results: In Codex, GPT-6 Sol (max) scores 57 on the Coding Agent Index, up 2 points over GPT-5.6, at roughly half the per-task cost; Luna (max) matches Sol High and Astra Low on DeepSWE at about one-third of Sol's price (relayed by Internet--Sensation).
- Lower hallucination: Artificial Analysis found AA-Omniscience hallucination rates falling from 92% to 60% for Sol and 93% to 77% for Luna (max effort).
- Knowledge-work regression: On GDPval-AA v2.1 (adapted from OpenAI's 44-occupation dataset), both Sol and Luna regress at max effort, mainly due to shorter answers omitting key points.
- DeepSWE decline: User tests (power97992, Angaisb and others) show GPT-6 Sol below GPT-5.6 Sol on DeepSWE; Luna improves slightly over 5.6 Luna at the same reduced price. kimmonismus's benchmark roundup (FrontierCode 48.4% xhigh) suggests Sol matches Fable 5.1 at 85% lower cost; krizzalicious49 posted benchmark screenshot collections versus the 5.6 series.
- AutomationBench highlight: reachvb relayed that Sol (xhigh) beats Opus 5 on AutomationBench at 9% of the cost.
- Positive hands-on: Dan Shipper of Every says GPT-6 Sol is now his daily Codex driver—writing near Astra quality, faster and 50% cheaper than 5.6 Sol.
Unconfirmed
- Rumors relayed by ns123abc claim Sol also trails 5.6 Sol on computer use and research debugging with no cybersecurity gains—unverified; commenters in kristoph's thread mocked it as a Terra rename, also rumor.
- sytelus claims Sol underperforms the previous generation on HealthBench and some bio/chem safety evals, without sources.
- weswinder and adonissingh suspect vendors deliberately "buried" the DeepSWE benchmark—speculation with no official response.
- power97992 reports Sol's xhigh losing to 5.6 Sol max on some coding tasks (small personal sample); FalconsArentReal relayed similar early user comparisons.
Why it matters
A new generation failing to beat its predecessor across the board is unusual, fueling debate about slowing generational progress and selective benchmark disclosure. Meanwhile, halved prices plus sharply reduced hallucination (Reddit users like indiantruely argue the pricing could squeeze many open- and closed-source models) make the models highly attractive for cost-sensitive production. With Anthropic and Qwen releasing the same day, the coding-API market is intensifying. Diverging signals between third-party benchmarks, user tests and daily-use impressions suggest choosing models by testing your own workloads rather than trusting "newer is better."
Timeline
- September 23: OpenAI announced GPT-6 Sol/Luna, live same day in ChatGPT Work, Codex and API; Altman posted multiple follow-ups on pricing and positioning; Artificial Analysis published its evaluation; Anthropic and Qwen also launched new products.
2026-09-23 ~ 2026-09-24 · 92 related posts
- Episode 1: OpenAI Launches GPT-6 Sol and Luna at Half Price; Big Hallucination Gains but Not a Clean Sweep(2026-09-23, 92 posts)
- Episode 2: GPT-6 Sol/Luna launch splits opinion: half-price but dumber?(2026-09-23, 9 posts)
- Episode 3: LMArena Opens GPT-6 Sol and Luna for Public Testing(2026-09-23, 2 posts)
- Episode 4: Four New Models Reshape the Cost-Performance Frontier, Opus 5.5 Tops Intelligence Index(2026-09-23, 2 posts)
- Episode 5: OpenAI and Anthropic Slash Prices Within 90 Minutes, Kicking Off a New Model Price War(2026-09-23, 7 posts)
- Episode 6: Tests Show GPT-6 Sol Scales Nearly Linearly With Effort(2026-09-23, 3 posts)
- Episode 7: GPT-6 Astra Tops Bug Hunt Bench While Sol Version Regresses(2026-09-24, 2 posts)
- Episode 8: Leaks Map GPT-6 Codenames to GPT-5.6 Versions(2026-09-24, 3 posts)
- Episode 9: Bug Hunt Bench Ranks Frontier Coding Models on 105 Real Bugs with 200x Cost Gap(2026-09-24, 3 posts)
- Episode 10: GPT-6 Sol and Luna Launch With Aggressive API Pricing(2026-09-25, 2 posts)
- Episode 11: OpenAI's efficient GPT-6 models seen as freeing compute for bigger frontier models(2026-09-25, 2 posts)
- Episode 12: GPT-6 Sol Underestimated: 1.6x Robot Task Gains at 47% Lower Cost(2026-09-25, 2 posts)
- Episode 13: GPT-6.1 Sol Launches with Astra-Level Intelligence at a Fifth of the Price(2026-09-29, 32 posts)
- Episode 14: GPT-6 Astra Ultrafast Reportedly 8x Faster in Codex(2026-09-30, 2 posts)
- Episode 15: GPT-6.1 Sol third-party reviews: near-Astra performance at a fraction of the cost(2026-09-30, 21 posts)
Primary sources
- GPT-6 Sol and Luna API Pricing Leaks: Sol Halves to $2/$10, Luna Drops to $0.10/$0.50 — Angaisb_ · 2026-09-23
- OpenAI launches GPT-6 family with Sol and Luna models — DemiPixel · 2026-09-23
- GPT-6 Sol and Luna are live: OpenAI ships new model pair — DemiPixel · 2026-09-23
- GPT-6 Sol and Luna pricing lands, Reddit says it could bury rival models — indian_truely · 2026-09-23
- GPT-6 Sol and GPT-6 Luna Officially Released as Dual Launch — soumitrashukla9 · 2026-09-23
- GPT 6 Sol and Luna pricing leak: 6 Sol at one-third the price of 5.6 — yeshvvanth · 2026-09-23
- GPT-6 Sol claimed to cost 50% less than GPT-5.6 Sol — cedric_chee · 2026-09-23
- GPT-6 Sol scores slightly below GPT-5.6 Sol on DeepSWE, only cheaper — Angaisb_ · 2026-09-23
- [source] GPT-6 Sol and Luna roll out in ChatGPT Work, Codex, and API — OpenAI · 2026-09-23
- OpenAI launches GPT-6 Sol and Luna: Sol beats Opus 5 at 9% of the cost — reach_vb · 2026-09-23
- OpenAI's Sol and Luna join GPT-6 family: better than GPT-5.6 at half the price — stevenheidel · 2026-09-23
- GPT-6 Sol and Luna launch with API prices 50% below GPT-5.6 — OpenAIDevs · 2026-09-23
- OpenAI credits caching and inference gains for GPT-6's 50% price cut — OpenAIDevs · 2026-09-23
- OpenAI launches GPT-6 Sol and Luna: near Fable 5 on DeepSWE at up to 96% lower cost — OpenAIDevs · 2026-09-23
- OpenAI launches GPT-6 Sol and Luna: faster models with 50% cheaper API — VraserX · 2026-09-23
- GPT-6 Sol leads AutomationBench over Fable 5.1 at 88% lower cost per task — OpenAIDevs · 2026-09-23
- GPT-6's improved prompt caching cuts agent costs; leads AutomationBench at 88% less — OpenAIDevs · 2026-09-23
- OpenAI says GPT-6 Sol/Luna improve on GPT-5.6 in alignment evals via Astra advances — OpenAIDevs · 2026-09-23
- GPT-6 Sol and Luna Priced at Half of GPT-5.6 Rates — No_Hovercraft6239 · 2026-09-23
- GPT-6 Sol and Luna benchmark screenshots surface on Reddit — krizzalicious49 · 2026-09-23
- GPT-6 benchmarks: Sol matches Fable 5.1 at ~85% less cost per task — kimmonismus · 2026-09-23
- Every's GPT-6 Sol Hands-On: Near-Astra Writing, 50% Cheaper Than 5.6 Sol — every · 2026-09-23
- GPT-6 Sol/Luna Benchmarks Leak: ~85% Cheaper Per Task Than Fable 5.1 — kimmonismus · 2026-09-23
- OpenAI Rolls Out GPT-6 Sol and Luna to All Paid ChatGPT Tiers Today — aziz4ai · 2026-09-23
- OpenAI: GPT-6 Sol beats Claude Opus 5 on AutomationBench at 9% of the cost — MaziyarPanahi · 2026-09-23
- GPT-6 Sol and Luna launch with strong coding at lower prices — cedric_chee · 2026-09-23
- [source] GPT-6 Sol and Luna Halve Prices, Sol Cuts Hallucination Rate from 92% to 60% — ArtificialAnlys · 2026-09-23
- GPT-6 Sol Gains 2 Points in Coding Agent Index at Half the Cost — ArtificialAnlys · 2026-09-23
- GPT-6 Models Hallucinate Substantially Less Than Predecessors — ArtificialAnlys · 2026-09-23
- GPT-6 Regresses on GDPval Knowledge-Work Benchmark Due to Shorter Deliverables — ArtificialAnlys · 2026-09-23
- GPT-6 Sol and Luna added to Artificial Analysis Intelligence Index v4.3 — ArtificialAnlys · 2026-09-23
- Artificial Analysis Breaks Down Individual Evals in Intelligence Index v4.3 — ArtificialAnlys · 2026-09-23
- 'Sir, Altman just dropped GPT-6 Sol at half the price of Opus 5.5' meme — ns123abc · 2026-09-23
- [source] Sam Altman announces GPT-6 Sol and Luna: major upgrades at half the price — sama · 2026-09-23
- OpenAI launches GPT-6 Sol and Luna, permanently cuts API prices 50% — aziz4ai · 2026-09-23
- Polymarket: GPT-6 claims up to 93% cheaper coding task costs than Claude Opus 5 — Polymarket · 2026-09-23
- Altman: GPT-6 Sol and Luna have no competition on per-task pricing — sama · 2026-09-23
- GPT-6 Luna Max matches Sol High on DeepSWE at 3x lower price — Internet--Sensation · 2026-09-23
- OpenAI posts GPT-6 Sol and Luna page with zero details — petrusenko_max · 2026-09-23
11 near-duplicate retellings: masterderptato · SteveEricJordan · minxio_ · Angaisb_ · kimmonismus · _philschmid · charliermarsh · polynoamial · gdb · Dimillian · 赛博禅心