GPT-6 Sol and Luna
TL;DR Highlight
GPT-6 Sol and Luna cut API prices 50% versus GPT-5.6, per HN comments.
Who Should Read
A backend developer running coding agents on the OpenAI API or Codex who is unsure which model to pick for cost.
Core Mechanics
- No article text, comments only
- The source body was not provided, so the article's own facts can't be verified.
- Everything below comes from comments.
- Comments quote a price table
- GPT-6 Sol is $4→$2 input and $20→$10 output versus GPT-5.6 Sol, 50% cheaper.
- GPT-6 Luna is $0.20→$0.10 input and $1.20→$0.50 output versus GPT-5.6 Luna.
Community Reactions
- Quoted: Luna's coding score went down
- A comment cites Artificial Analysis scoring GPT-6 Luna (max) 41 on the Coding Agent Index, 2 points below GPT-5.6 Luna (max).
- Cost per task is reportedly about 60% lower.
- Quoted: Sol improved on score and cost
- A comment says GPT-6 Sol (max) scores 57, up 2 points from GPT-5.6 Sol (max).
- It reportedly costs $2.99 per task, about 50% less.
- Comment says subscription usage doubles too
- A commenter linked a source saying the price cut applies to subscriptions, not just the API.
- Complaint that model and effort choices are too complex
- Commenters say the Astra, Sol and Luna lineup plus effort levels make per-task selection hard.
- Experience moving off fine-tuned gpt-4.1-mini
- A commenter said gpt-5-mini was less accurate, slower and pricier, but GPT-6 Luna matches their fine-tuned accuracy at a competitive price.
How to Apply
- Compare current model costs in a table
- Pull monthly token volume for your GPT-5.6 Sol and Luna calls and recompute cost with the comment's price table.
- Confirm the numbers on the official pricing page before deciding.
- Run Luna 6 against 5.6 on your own evals
- Run 20-50 simple refactoring or structured-output tasks on both models with the same prompt.
- Record accuracy and cost per task together.
Try It
# Paste into ChatGPT or Claude (model selection check prompt)
Here is the list of LLM tasks in our service.
{{task list: name, monthly calls, avg input/output tokens, required accuracy}}
Price table (source: HN comments, verify on the official page, per 1M tokens):
- GPT-6 Sol: input $2, output $10
- GPT-6 Luna: input $0.10, output $0.50
Compute monthly cost per task, then split which tasks to try on Luna first and which need Sol.
Next: ask it to propose three evaluation criteria for each task to be tested.Terminology
Related Papers
Mercury 3: Diffusion at scale
Scaling diffusion LLM 40x gives Mercury 3: 4–8x faster than speed-tier models, similar quality
Timestral: Mistral's time-series foundation models
A 209M time-series model ranks first on three benchmarks and runs on CPU.
DeepSeek V4.1 Flash 1017 (Max)
DeepSeek V4.1 Flash 1017 Max scores 45 on the Intelligence Index at 211 tok/s and $0.30 input.
If you don't run your own LLM proxy, how do you know whats going on?
Running your own LLM proxy shows which keys, models, and cache misses drive your usage.
DeepSeek: DeepSeek v4.1 Flash 1017
DeepSeek V4.1 Flash 1017: $0.05/M input, 1M context, built for agents
DeepSeek-v4.1-Flash-1017
DeepSeek V4.1 Flash minor update lifts coding and agent benchmarks while cutting run time and output length
Show HN: Llmspeed – feel how fast a local LLM runs before you buy the hardware