Best value AI model
GPT-6 Luna is the best value: about $0.31 per 1,000 correct coding answers over the last 30 days, every attempt billed, the lowest among the 14 models within five points of the best on the Coding board in the last 24 hours.
model: auto-value · your own provider keys · 1,000 requests a month free
Free account · 24 hours, 7 days and 30 days · no card
Already have an account? Log in
As of 26 Oct 2026, 11:22 UTC · last 30 days
- $0.31
- GPT-6 Luna: per 1,000 correct coding answers, every attempt billed, last 30 days
- 14
- models within five points of the best on the Coding board in the last 24 hours
- 92%
- of GPT-6 Luna’s coding attempts passed the hidden tests over the window
01
Which AI model is the best value over the last 30 days?
GPT-6 Luna is the best value: about $0.31 per 1,000 correct coding answers over the last 30 days, every attempt billed, the lowest among the 14 models within five points of the best on the Coding board in the last 24 hours.
Cost here is what a model’s attempts cost, not its list price: every model runs the same coding tasks, every attempt is billed, and the total is divided by the attempts that passed the hidden tests. Only models within five points of the best on the Coding board in the last 24 hours qualify, so value never means cheaper and much worse.
| Model | Vendor | Coding score (24 h) | Pass rate (30 days) | Cost per 1,000 correct answers | Median response |
|---|---|---|---|---|---|
| GPT-6 Luna | OpenAI | 92 (=11) | 92% | $0.31 | 3.5 s |
| GPT-5.6 Luna | OpenAI | 95 (=1) | 90% | $0.78 | 3.3 s |
| DeepSeek V4.1 Flash | DeepSeek | 96 (=1) | 90% | $0.82 | 9.3 s |
| Gemini 3.8 Flash | 91 (=11) | 90% | $4.89 | 4.7 s | |
| Claude Sonnet 5.5 | Anthropic | 93 (=5) | 86% | $6.25 | 10 s |
Free account: every ranking over 24 hours, 7 days and 30 days, a 3-model watchlist with email alerts, 1,000 routed requests a month and a Data API key. No card.
Already have an account? Log in
02
How the leader is decided
The leaderboard’s Best value card has one stated definition: The most points per measured dollar of one identical coding run — every attempt billed — among the models within five points of the top.
This page applies the same bar — within five points of the best, then the lowest cost per correct answer — over the last 30 days, so a model that is cheap and much worse never appears. Cost per correct answer counts every attempt, failed ones included.
On the Coding board in the last 24 hours the top score is 96 (DeepSeek V4.1 Flash), so the bar is 91: 14 models clear it, and this page orders them by cost per correct answer over the last 30 days, from the same runs as the scores.
The score column is the Coding board’s own, with its statistical ties written “=”; the definitions are on the methodology page.
03
What the cost per correct answer measures
Cost is measured, not list price: every model runs the same seven coding tasks and every attempt is billed, so a verbose model costs what it really costs per task.
- Cost per correct answer
- What all of a model’s attempts cost over the last 30 days, failed ones included, divided by the attempts that passed the hidden tests. Shown per 1,000 correct answers.
- Pass rate
- The share of attempts whose hidden tests passed over the same window, shown beside the cost.
- The bar
- Only models within five points of the best on the Coding board in the last 24 hours qualify: the rule the leaderboard’s Best value card and the Smart Router’s value modes share, so value never means cheaper and much worse.
- List price
- Not used here. List price per million tokens is on every head-to-head page; a model that writes four times as much costs four times as much per task at the same list price.
The methodology is public — every weight, statistical test and drift constant. The test tasks are private, so no lab can train on them. The full method is on the methodology page.
04
Does the best value model change week to week?
Sometimes. The order moves when a provider changes a price, a model gets faster or slower, or a pass rate shifts, and the window is the last 30 days so a single day cannot decide it on its own. The bar moves with the Coding board, which re-ranks after every run: coding every four hours, reasoning and tool use daily, with an hourly canary probe.
A model that slips for good is a different question: every suite has its own Page-Hinkley (a CUSUM-family test) change-point detector that compares each model with its own past, and the current answer, model by model, is on Is AI getting worse?. If the question is whether ChatGPT or Claude is getting worse, or a model you use has been nerfed, the family pages answer it with the same data (is ChatGPT getting worse, is Claude getting worse); the method is on AI drift detection.
Three models free · email when a watched model's coding score falls five points in a week
05
Is Claude better value than GPT?
Among the models within five points of the best on the Coding board in the last 24 hours, the cheapest Claude model is Claude Sonnet 5.5 at $6.25 per 1,000 correct coding answers and the cheapest GPT model is GPT-6 Luna at $0.31 per 1,000 correct coding answers, over the last 30 days. GPT is about 20× cheaper per correct answer in this window.
Both on one page: Claude Sonnet 5.5 vs GPT-6 Luna. Every Claude model is on Anthropic models and every GPT model on OpenAI models; the other four families — Google (Gemini), DeepSeek, Moonshot AI (Kimi) and Z.ai (GLM) — are on the same board above.
06
What about speed?
Of the same group, the fastest is GPT-5.6 Luna, a median of 3.3 s per coding attempt over the last 30 days. The ranking by response time is on fastest AI model.
The board pages rank ability alone: best LLM for coding, best LLM for reasoning and best AI model for agents. The head-to-head pages put list price, cost per correct answer and response time beside the scores.
07
How do I route to it automatically?
One OpenAI-compatible endpoint sends each request to whichever model the boards favour for that kind of work, with your own provider keys.
Set model to auto-value on the Smart Router’s OpenAI-compatible endpoint and each request goes to the most combined-score points per measured dollar among models within five points of the best one you can use. auto-value-coding picks the same rule on the Coding board alone, for coding traffic. The router ranks on the same benchmark results as this page and re-decides as they move; routing is drift-aware by default, uses your own provider keys and adds zero markup on tokens. Free includes 1,000 routed requests a month.
Every auto-* mode is described in the router documentation.
1,000 routed requests a month free · your own provider keys · no card
08
Best value AI model: questions people ask
Which AI model is the best value in October 2026?
GPT-6 Luna is the best value: about $0.31 per 1,000 correct coding answers over the last 30 days, every attempt billed, the lowest among the 14 models within five points of the best on the Coding board in the last 24 hours. As of 26 Oct 2026, 11:22 UTC. The window rolls forward every day, so the answer on this page follows the price view.
Is Claude better value than GPT?
Among the models within five points of the best on the Coding board in the last 24 hours, the cheapest Claude model is Claude Sonnet 5.5 at $6.25 per 1,000 correct coding answers and the cheapest GPT model is GPT-6 Luna at $0.31 per 1,000 correct coding answers, over the last 30 days. GPT is about 20× cheaper per correct answer in this window. Both on one page: Claude Sonnet 5.5 vs GPT-6 Luna.
How is the best value model decided?
The leaderboard’s Best value card has one stated definition: The most points per measured dollar of one identical coding run — every attempt billed — among the models within five points of the top. This page applies the same bar — within five points of the best, then the lowest cost per correct answer — over the last 30 days. On the Coding board in the last 24 hours the top score is 96 (DeepSeek V4.1 Flash), so the bar is 91: 14 models clear it, and this page orders them by cost per correct answer over the last 30 days, from the same runs as the scores. The definitions are on the methodology page.
Is the cheapest model actually cheaper after retries?
Not always, which is why this page counts every attempt. Cost per correct answer divides what all of a model’s attempts cost over the last 30 days, failed ones included, by the attempts that passed the hidden tests, so a model that fails often pays for its retries. GPT-6 Luna is the lowest at about $0.31 per 1,000 correct coding answers, with a pass rate of 92%. List price per million tokens is a different number and is shown on every head-to-head page.
How often does the ranking update?
Every model on the board runs the same tasks on the same schedule: coding every four hours, reasoning and tool use daily, with an hourly canary probe. This page reads the price view’s last 30 days and refreshes every fifteen minutes. The methodology is public — every weight, statistical test and drift constant. The test tasks are private, so no lab can train on them.
Can I route to the best value model automatically?
Set model to auto-value on the Smart Router’s OpenAI-compatible endpoint and each request goes to the most combined-score points per measured dollar among models within five points of the best one you can use. auto-value-coding picks the same rule on the Coding board alone, for coding traffic. The router ranks on the same benchmark results as this page and re-decides as they move; routing is drift-aware by default, uses your own provider keys and adds zero markup on tokens. Free includes 1,000 routed requests a month. Every auto-* mode is described in the router documentation.
Route to the best automatically
Free account: every ranking over 24 hours, 7 days and 30 days, a 3-model watchlist with email alerts, 1,000 routed requests a month and a Data API key. No card.
model: auto-value · your own provider keys · zero markup on tokens
Free account · 24 hours, 7 days and 30 days · no card
Already have an account? Log in