OpenAI and Anthropic trade blows in AI price war: GPT-6 Sol/Luna vs Claude Opus 5.5
OpenAI halved API prices for GPT-6 Sol and Luna within an hour of Anthropic cutting Opus 5.5 prices 20%, intensifying the model pricing race.
What to know
- Anthropic and OpenAI released competing model updates within about an hour of each other on September 22, 2026, both featuring major price cuts.
- GPT-6 Sol and Luna are priced roughly 50% below their GPT-5.6 predecessors; Claude Opus 5.5 got a 20% cut with a 60% drop in cache-read pricing.
- Independent Artificial Analysis benchmarks show real tradeoffs: Sol improved on coding tasks while Luna regressed slightly, and both cut hallucination rates but lost ground on some knowledge-work evaluations.
- Developer discussion centers less on raw token prices and more on practical usage limits, caching mechanics, and whether Claude can remain competitive against OpenAI's cheaper tiers.
The dispute Whether the halved token prices represent a straightforward win or are offset by benchmark regressions and confusing real-world usage-limit mechanics. · positions read across 74 posts and comments
The price cuts are a huge, potentially decisive competitive move that makes OpenAI's models hard to pass up.
-
“GPT-6 Luna being half the price of GPT-5.6 Luna is a really big deal.”
simonw · Hacker News ↗
Benchmark regressions (especially Luna's coding drop) show the price cuts aren't a free lunch.
-
“Artificial Analysis is reporting that 6 Luna scores 2 points lower on their coding index than 5.6 Luna, but is 60% cheaper”
scrlk · Hacker News ↗
What matters day-to-day is opaque usage limits and plan mechanics, not the headline API price.
-
“At this point, the deciding factors for me between Claude Code 20x and Codex Pro 20x are... usage limits: downstream of input/output cost, but resets and obscure windows and odd 20x plan / 5x plan != 4x usage math throw a wrench into it.”
jeffnash · Hacker News ↗
OpenAI Developer of GPT-6 Sol and LunaAnthropic Developer of Claude Opus 5.5Simon Willison Independent AI commentator/bloggerThariq Shihipar Commentator quoted on Opus 5.5Artificial Analysis Independent AI benchmarking firm
How it unfolded 6 developments, newest first · click a bar or a number to jump articlespostscomments
-
6
Coverage shifts focus to caching as the real cost lever
Follow-up reporting argued that prompt-caching discounts, not headline token prices alone, are the bigger factor driving down real-world inference costs for GPT-6 Sol and Luna.
“It communicates clearly, it's cheaper per token than Opus 5.0 with the intelligence of Fable 5.1 it's very token efficient and works across every effort level.”
— Thariq Shihipar, Quoted commentator · source -
first by The New Stack, 19h ago
-
> This year, coding agents have begun tackling tasks with more complexity, scope, and duration than ever before. At OpenAI, our internal usage has grown exponentially. Valued at API prices, daily token usage has exceeded $600 for the median researcher and $7,000 for researchers at the 90th percentile (Research acceleration: The view inside OpenAI…
2 more of the top 3 · 14 posts in this stretch
-
A
OpenAI has introduced GPT-6 Sol and Luna, new coding-focused models with lower API prices, benchmark claims, and staged ChatGPT access https:// alternativeto.net/news/2026/9/ openai-introduces-gpt-6-sol-and-luna-with-lower-api-prices/
-
J
#OpenAI launches #GPT-6Sol and #Luna, boasting lower cost and fewer mistakes
-
-
5
Willison's hands-on tests find Opus 5.5 fails its pelican benchmark
Simon Willison's informal SVG-pelican test showed Claude Opus 5.5 at max thinking failing to return a result for the first time, while GPT-6 Luna became one of the cheapest models OpenAI has ever shipped and GPT-5.6 Terra lost its remaining reason to exist once priced level with Sol.
-
first by Tech Times, 1d ago · also Digit, Tech in Asia, The GitHub Blog, The Neuron, Constellation Research, Wccftech +6
12 more headlines
- OpenAI launches GPT 6 Sol and Luna with Astra-level AI at lower cost, check details Digit · 1d ago
- OpenAI launches GPT-6 Sol, Luna at half the API cost Tech in Asia · 1d ago
- OpenAI's GPT-6 Sol and GPT-6 Luna now available The GitHub Blog · 1d ago
- 😺 GPT-6 Sol / Luna vs. Claude Opus 5.5 LIVE The Neuron · 1d ago
- OpenAI adds GPT-6 Luna and Sol, touts lower prices Constellation Research · 1d ago
- OpenAI Introduces GPT-6 Sol and Luna With 50% Lower API Prices Unite.AI · 1d ago
- OpenAI Unleashes A New Price War, With GPT-6 Sol And GPT-6 Luna Now Priced Below Claude Opus 5.5 And DeepSeek's V4.1 Flash, Respectively, Negating The Rationale For Open-Weight Models Wccftech · 1d ago
- OpenAI launches GPT-6 Sol and Luna with improved intelligence at lower prices Neowin · 1d ago
- OpenAI Launches GPT-6 Sol And Luna With 50% Lower API Pricing Pulse 2.0 · 1d ago
- OpenAI launches GPT-6 Sol and Luna with sharply lower API prices RuntimeWire · 1d ago
- Introducing GPT-6 Sol and Luna OpenAI · 1d ago
- OpenAI Launches GPT-6 Sol and Luna Minutes After Anthropic Drops Claude Opus 5.5 Decrypt · 1d ago
-
first by Unite.AI, 1d ago
-
C
📰 #Codex Daily Brief - 2026/09/23 🔥 今日のトップ3 1. GPT-6 Sol / Luna が Codex に投入、API価格は半額 - 複雑コーディングはSol、量産作業はLuna。トークン単価は前世代の半分 - 重要度：★★★★★ - 出典 openai.com/index/introd... techcrunch.com/2026/09/22/o... learn.chatgpt.com/docs/changelog 👇️続く
2 more of the top 3 · 13 posts in this stretch
-
So with GPT-6 Astra, Codex introduced an experimental feature for context management that’s supported to be beneficial for long conversations. Will that experimental feature now also apply to Sol and Luna?https://community.openai.com/t/experimental-context-manageme...I would also like to point out that it was quite predictable that Terra got…
-
Big model release today - I wrote about Claude Opus 5.5, GPT-6 Sol, and GPT-6 Luna - plus comparison grids of pelicans by the different model families at different reasoning levels https:// simonwillison.net/2026/Sep/22/ opus-and-sol-and-luna/
-
-
4
Artificial Analysis benchmarks show mixed gains for Sol and Luna
Sol improved 2 points on the Coding Agent Index while Luna regressed 2 points; both models cut hallucination rates sharply (Sol from 92% to 60%) but showed regressions on knowledge-work evaluations like GDPval-AA.
“GPT-6 Sol (max) gains 2 points in the Artificial Analysis Coding Agent Index at half the Cost per Task of its predecessor, while GPT-6 Luna (max) regresses by 2 points.”
— Artificial Analysis -
first by Mashable, 1d ago
-
I wish there was more transparency on the plus plans usage limits showing actual token usage and prices per model that eats away at remaining usage.Does anyone know how exactly these price differences for example between sol6 and sol5.6 translate to codex percentages? In theory it seems like for "high" on both it should result in ~3x more usage…
2 more of the top 3 · 13 posts in this stretch
-
Simon, love your work, one piece of minor feedback for the individual model pages is to make the font of the model name potentially bigger than (and above) the conversation id (which means nothing to the audience) "2026-09-22T18:28:00 conversation: 01m355zvyw8946qyraa8zpz6h9 id: 01m355zvyx47zxx5c6q6b3fg0m#".I had all the tabs open individually and…
-
From the perspective of “an average person”, ChatGPT is delivering fantastic products.- For general chat and web search, occasional image editing, small coding work, document review etc. ChatGPT Plus is basically limitless and “just works” since 5.6. I’ve yet to give it some task it cannot do.- When given sensible instructions, it hardly annoys…
-
-
3
Developers react to the price cuts and usage-limit tradeoffs
On Hacker News, commenters called the Luna price cut a major deal while others weighed API savings against opaque ChatGPT plan usage limits and Codex quirks.
“GPT-6 Luna being half the price of GPT-5.6 Luna is a really big deal.”
— simonw -
first by Business Today, 1d ago · also MarkTechPost, Gizmodo, iPhone in Canada, VentureBeat, The New Stack
5 more headlines
- OpenAI Releases GPT-6 Sol and Luna: 50% Cheaper API Pricing and Benchmarks MarkTechPost · 1d ago
- OpenAI Releases New GPT-6 Sol and Luna Models, Sells Them as Mini-Astras Gizmodo · 1d ago
- OpenAI's New GPT-6 Sol and Luna Models Are 50% Cheaper iPhone in Canada · 1d ago
- OpenAI releases GPT-6 Sol and Luna models, slashing API costs 50% or more VentureBeat · 1d ago
- OpenAI releases GPT-6 Sol and Luna — and cuts token prices in half The New Stack · 1d ago
-
GPT-6 Sol and Luna are out. Not only are they a very significant improvement across the board, but also in writing and general "you know when you try it" quality. We are also permanently reducing the API price by 50% making both of them viable for a ton of new usecases and
2 more of the top 3 · 19 posts in this stretch
-
At this point, the deciding factors for me between Claude Code 20x and Codex Pro 20x are:1/ Usage limits: downstream of input/output cost, but resets and obscure windows and odd 20x plan / 5x plan != 4x usage math throw a wrench into it. Winner right now is Codex by a mile, especially when you factor in ChatGPT usage (even 6 Astra Pro) is…
-
N
GPT-6 Sol and Luna: https:// openai.com/index/introducing-g pt-6-sol-and-luna/ Discussion: http:// news.ycombinator.com/item?id=4 9805509
-
-
2
OpenAI launches GPT-6 Sol and Luna at half GPT-5.6 prices
OpenAI released GPT-6 Sol and Luna, extending Astra's improvements to cheaper tiers, and cut API prices roughly 50% versus GPT-5.6 (Sol from $4/$20 to $2/$10; Luna from $0.20/$1.20 to $0.10/$0.50).
“GPT‑6 Sol and Luna build on Astra's advances in alignment, showing improvements over their GPT-5.6 counterparts.”
— OpenAI -
10 outlets first by The Verge, 1d ago · also HN Best, HN Frontpage, ZDNET, Bitcoin Insider, iThinkDifferent, TestingCatalog AI News +3 · read ↗
-
first by Indian Express, 1d ago · also The Economic Times, Reuters, Channel News Asia
-
GPT‑6 Sol and Luna build on Astra’s advances in alignment, showing improvements over their GPT-5.6 counterparts.
2 more of the top 3 · 15 posts in this stretch
-
I don't see how anyone can be using Claude with prices like this, it's pretty incredible what the OpenAI team is doing, w.r.t model quality and pricing. Prices per 1M tokens Claude Opus 5.5 Claude Opus 5 Cache reads $0.20 $0.50 Input tokens $4 $5 Output tokens $20 $25 Cache writes $5 $6.25 ModelInputOutputPrice reductionGPT‑6 Sol vs. GPT‑5.6 Sol$4…
-
T
OpenAI is launching two new models, which it says are cut from the same cloth as Astra. https:// techcrunch.com/2026/09/22/open ai-launches-gpt-6-sol-and-luna/?utm_source=dlvr.it&utm_medium=mastodon
-
-
1
Anthropic releases Claude Opus 5.5 with a 20% price cut
Opus 5.5 dropped from $5/$25 per million input/output tokens to $4/$20, with cache-read prices falling 60%, addressing complaints about Opus's communication style.
“Anthropic released a new version of Opus 5.5 just 90 minutes before OpenAI's release, reflecting the intense competition between the two companies.”
— TechCrunch
What people are saying 12 voices from 3 sites · best of 74 · verbatim
- How exactly do the API price cuts translate into ChatGPT plan usage-limit percentages (e.g., Codex 'high' mode)?
- Will the experimental context-management/compaction feature introduced with Astra also apply to Sol and Luna?
- Can Anthropic's upcoming Sonnet 5.5 and Haiku 5.5 regain price competitiveness against GPT-6 Luna at the low end?
- Sep 22
-
It seems I'm one of the few Terra users since Astra dropped?When 5.6 dropped I had no weekly limits and I could just drive my work with Sol xhigh and things were great. Once limits were back (and maybe token prices changed iirc) Sol was no longer usable (on Pro or business) unless I was ok with 4 prompts every 5 hours, so I had to switch to Terra…
-
GPT-6 Luna being half the price of GPT-5.6 Luna is a really big deal.Here's GPT-6 Luna pelicans: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...And GPT-6 Sol: https://tools.simonwillison.net/markdown-svg-renderer?url=ht...Scroll to the bottom for the GPT-6 Sol max one…
-
> It's also is a godsend if you use an MCP like oracle to automate the process of calling the Pro model on particularly tough problemsAs far as I know Codex (at least the GUI) can automatically call the ChatGPT Chat models (including Astra 6 Pro), you just need to @ a ChatGPT Chat conversation from within Codex and tell it when to use it.> There…
-
Artificial Analysis is reporting that 6 Luna scores 2 points lower on their coding index than 5.6 Luna, but is 60% cheaper:> In the Coding Agent Index, Sol improves but Luna regresses: In OpenAI's Codex harness, GPT-6 Sol (max) scores 57 in the Artificial Analysis Coding Agent Index, up 2 points from GPT-5.6 Sol (max), with gains in Terminal-Bench…
-
G
https:// openai.com/index/introducing-g pt-6-sol-and-luna/ 這真熱鬧，OpenAI 也更新了 Sol & Luna 產品線...
-
🚨 GPT-6 Sol is HALF the price of GPT-5.6 Sol. Standard API pricing per 1M tokens (input / output): Sol: $4 / $20 → $2 / $10 Luna: $0.20 / $1.20 → $0.10 / $0.50 That's 50% cheaper input and nearly 60% cheaper output for Luna.
-
It seems OpenAI are making up for 6 Sol likely not being as good as Opus 5.5 through its pricing. Still waiting for benchmarks to drop, but... GPT-6 Sol is 50% of the price of 5.6 Sol. $2 in per million tokens, $10 out. 🤑
-
GPT-6 Sol and Luna just landed in Astra's orbit. Both launch today with API prices 50% lower than GPT-5.6. Build with Sol. Scale with Luna. To production and beyond.
-
It's a double release day just like old times! GPT-6 Sol and GPT-6 Luna are rolling out, for some of my mutuals they are already live. OpenAI announcement incoming.
-
GPT-6 Sol and Luna are dirt cheap GPT-6 Sol is half the price of Opus 5.5
-
GPT-6 Sol and Luna benchmarks from the official blogpost!
-
H
GPT-6 Sol and Luna L: https:// openai.com/index/introducing-g pt-6-sol-and-luna/ C: https:// news.ycombinator.com/item?id=4 9805509 posted on 2026.09.22 at 14:00:34 (c=0, p=6)