LLM Releases, Token Limits, and Pricing Since GPT 3.5
Comments on “LLM Releases, Token Limits, and Pricing Since GPT 3.5”
Claim-checked. Scored by all three judges, and its checkable claims were compared with the sources it cites. How far stories are checked
Engineer Mete Morris compiled launch data for proprietary models from five US labs, grouped into economy, balanced, flagship and premium tiers. He reports median flagship release gaps at OpenAI, Anthropic and Google falling from 119 days in 2023 to 57 in 2026, and median flagship output limits rising from about 4K to 128K tokens. Prompt capacity rose from 100K to 1M. Launch prices vary by tier, and the output-to-input price ratio shifts with thinking settings. Several figures could not be independently checked, and prices are list prices at launch only.
Why it ranked
No. 9 of 60 with a score of 0.68. The panel rated it highest on new information (0.85) and audience breadth (0.81), lowest on significance (0.58).
- AI judgeJudge 2claude-haiku-5-5
- Data-driven analysis of LLM release cadence, context and output limits, and pricing trends, useful to builders. Sourcing is mostly embedded charts, and some claims are interpretive guesses, so substantiation is moderate.
- AI judgeJudge 3gpt-5.6-luna
- A data-backed analysis of LLM release cadence, context/output limits, and pricing offers timely, broadly useful AI-market and infrastructure insight. It provides methodology and comparisons, though some causal interpretations remain speculative.
Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.
Score
0.68total under rubric v2.3
- Significance0.58
- Reader interest0.77
- Audience breadth0.81
- Practical relevance0.71
- Substantiated0.73
- Technical depth0.65
- New information0.85
- Counts against: Promotional0.05
- Counts against: Overclaim0.22
Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).
| Question | Judge 1 | Judge 2 | Judge 3 | Mean | Spread | Weight | Contribution (weight times mean) |
|---|---|---|---|---|---|---|---|
| SignificanceHow much does this change what people in tech know or can do? | 0.68 | 0.40 | 0.67 | 0.58 | 0.28 | +0.24 | +0.140 |
| Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on? | 0.70 | 0.72 | 0.90 | 0.77 | 0.20 | +0.26 | +0.201 |
| Audience breadthHow many people who build technology does it concern? | 0.90 | 0.70 | 0.82 | 0.81 | 0.20 | +0.13 | +0.105 |
| Practical relevanceHow useful is it to someone who builds software or tech products? | 0.81 | 0.60 | 0.72 | 0.71 | 0.21 | +0.22 | +0.156 |
| SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations. | 0.82 | 0.60 | 0.78 | 0.73 | 0.22 | +0.05 | +0.037 |
| Technical depthHow technically substantive is it? | 0.78 | 0.50 | 0.68 | 0.65 | 0.28 | +0.05 | +0.033 |
| New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)? | 0.89 | 0.70 | 0.96 | 0.85 | 0.26 | +0.05 | +0.043 |
| PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional. | 0.07 | 0.05 | 0.03 | 0.05 | 0.04 | −0.19 | −0.010 |
| OverclaimDo the headline or conclusions claim more than the body supports? | 0.19 | 0.30 | 0.18 | 0.22 | 0.12 | −0.11 | −0.025 |
| Sum of contributions | +0.680 | ||||||
| Claim-check penalty | 0.000 | ||||||
| Total, as published | 0.680 | ||||||
Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.
The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.
<a href="https://rubric.hrih.si/editions/2026-11-16/9/"><img src="https://rubric.hrih.si/badge/2026-11-16/9.svg" alt="Ranked #9 on Rubric, 16 Nov 2026" width="238" height="28"></a>Claim check
Claim check: 3 matches source · 5 couldn’t check. Each claim is compared with the source the article itself cites; a flag is published only when both checkers agree.
Couldn’t checkDate
ChatGPT launched in November 2022, powered by GPT-3.5.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Since ChatGPT launched in November 2022, powered by GPT-3.5”
The source says No passage could be compared.
Source cited by the article: openai.com/index/chatgpt
Couldn’t checkNumber
Google offered a 2M-token context window with Gemini 1.5 Pro, and no release since 2025 has approached that.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “Google offered a 2M-token context window with Gemini 1.5 Pro but no release has approached that since 2025”
The source says No passage could be compared.
Source cited by the article: developers.googleblog.com/en/new-features-for-the-gemini-api-and-google-ai-studio
Couldn’t checkNumber
For OpenAI, Anthropic, and Google, the median flagship release gap fell from 119 days in 2023 to 47.5 days in 2025, then rose to 57 days in 2026.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “the median fell from 119 days in 2023 to 47.5 days in 2025. It rose to 57 days in 2026”
The source says No passage could be compared.
Source cited by the article: metemorris.com/posts/tokenomics/data
Couldn’t checkNumber
For flagship models, 10 of the 12 context windows for 2026 releases were about 1M tokens, up from 4 of 16 in 2025.
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “10 of the 12 context windows for 2026 releases were about 1M tokens, up from 4 of 16 in 2025”
The source says No passage could be compared.
Source cited by the article: metemorris.com/posts/tokenomics/data
Matches sourceNumber
GPT-6 Luna and Claude Haiku 5.5 both launched at $0.10 input and $0.50 output per million tokens.
The cited source says the same thing.
The article says “GPT-6 Luna and Haiku 5.5 both launched at $0.10 input and $0.50 output per million tokens”
The source says “gpt-6-luna | OpenAI | Economy | 2026-09-22 | 922,000 | 128,000 | 1,050,000 | $0.10 | $0.50 | Release”
Source cited by the article: metemorris.com/posts/tokenomics/data
Matches sourceNumber
GPT-4 launched at $30 input and $60 output per million tokens, while GPT-5 launched at $1.25 and $10.
The cited source says the same thing.
The article says “GPT-4 launched at $30 input and $60 output per million tokens (2×). GPT-5 launched at $1.25 and $10 (8×”
The source says “gpt-4-0314 | OpenAI | Flagship | 2023-03-14 | Shared window | — | 8,192 | $30.00 | $60.00 | Release”
Source cited by the article: metemorris.com/posts/tokenomics/data
Matches sourceBenchmark result
The authors of DFlash reported more than 6× faster generation than regular decoding in their benchmarks.
The cited source says the same thing.
The article says “The authors reported more than 6× faster generation than regular decoding in their benchmarks”
The source says “Experiments show that DFlash achieves over 6 \times lossless acceleration across a range of models and tasks, delivering up to 2.5 \times higher speedup than the state-of-the-art speculative decoding method…”
Source cited by the article: arxiv.org/abs/2602.06036
Couldn’t checkNumber
For flagship models, the median output limit rose from about 4K tokens in 2023 to 128K in 2026 (about 31×), and the median prompt capacity from 100K to 1M (10×).
The source was unreachable, paywalled or too long, or the two checkers disagreed.
The article says “the median output limit went from about 4K tokens in 2023 to 128K in 2026, about 31×, and the median prompt capacity went from 100K to 1M, 10×”
The source says No passage could be compared.
Source cited by the article: metemorris.com/posts/tokenomics/data
Also covered by
No other coverage was clustered with this story.
Elsewhere
Elsewhere: HN 1 pts, not on front page
Recorded for context only, checked 16 Nov 2026, 05:47 UTC. It played no part in the ranking.