Model Comparison
GPT-5.6 Sol vs GPT-6 Astra
Too close to call: the best configurations, GPT-5.6 Sol (low) and GPT-6 Astra (max), are 33 Elo apart, and their 95% ranges overlap. Among 78 models, each ranked by its best configuration, GPT-5.6 Sol is #8 and GPT-6 Astra is #11. Cost per script: $0.113 for GPT-5.6 Sol (low) and $1.15 for GPT-6 Astra (max).
Last updated , from board data generated at 01:13 UTC.
Picks and Verdict
PickGPT-5.6 Sol
GPT-5.6 Sol (low)
Pick GPT-5.6 Sol (low) for the bulk of your scripts, because at $0.113 and about a minute the board can't separate it from Astra's best.
PickGPT-6 Astra
GPT-6 Astra (max)
Pick GPT-6 Astra (max) when you want the least slop and the closest length, with a slop score about three points above GPT-5.6 Sol's best, at $1.15 a script.
The setting that decides this page for me is the cheapest one on it. GPT-5.6 Sol (low) costs $0.113 a script and takes about a minute, and it is also the top setting here, 33 Elo above GPT-6 Astra (max), which costs about ten times as much and takes more than nine minutes. Their 95% ranges overlap, so the board can't separate the two. Ranked by best setting, GPT-5.6 Sol is the #8 model of 78 and GPT-6 Astra the #11 model. More thinking does nothing for Sol: ultra rates 5 Elo below low, for more than three times the price. Ultra adds cost and nothing else. Astra goes the other way. It gains 182 Elo from medium to max, and GPT-6 Astra (medium) falls 215 Elo below Sol at low, with separate ranges, while costing more than twice as much. Astra's best does earn something: about three points better on the slop score and closer to the requested length. Still, I'd spend $0.113 on GPT-5.6 Sol at low before $1.15 on Astra.
Disclosure: both models come from the same lab as one of the four judges, GPT-6.1 Sol. The other judges sit outside that lab, so neither is scored by its own lab alone (how the panel works). The judge analysis measures how much each judge favors its own lab.
How to Read the Writing Score, Writing Elo and 95% Ranges
Writing Score and Writing Elo. The Writing Score (0 to 100) is the weighted average of the nine metric scores: seven from the judges, Length Adherence from code and the Slop Score half from each. The Writing Elo is a rating from pairwise comparisons: on each script, each configuration's average across its drafts is compared with every other configuration's, and gaps within the draft-to-draft variation count as draws. Elo rewards winning often, so the two can put configurations in a different order.
Elo ranges. Each 95% range shows how far a configuration's Elo could move if the drafts had come out slightly differently. It comes from resampling every configuration's recorded draft scores on each reference script and ranking again. Comparing two separate ranges does not test the Elo gap between them: ranges far apart are strong evidence for the order, and overlapping ranges do not mean a tie. The reference scripts stay fixed, so the ranges say nothing about new scripts. The methodology page has the full description.
All Thinking Levels of Both Models
Every ranked configuration of GPT-5.6 Sol and GPT-6 Astra, scored on the same 10 scripts by the same four judges.
- Each row is one configuration: the model at one thinking level. The dot is its Writing Elo, a rating from pairwise comparisons of the scores on each script, and the bar is its 95% range: how far that Elo could move if the drafts had come out slightly differently.
- When two bars overlap, the board cannot order those two configurations. Bars far apart are strong evidence for the order.
- The table under the chart lists the same configurations in the same order and adds the Writing Score (the 0 to 100 consensus of the judges and code), cost, time and configuration rank (the place among all configurations on the board).
Shaded rows are the configurations chosen for the metric comparison below.
| Configuration | Writing Elo | 95% Range | Writing Score | Cost / Script | Time / Script | Configuration Rank |
|---|---|---|---|---|---|---|
| GPT-5.6 Sol (low) | 2218 | 2181 to 2251 | 82.3 | $0.113 | 0.9 min | #16 |
| GPT-5.6 Sol (ultra) | 2213 | 2173 to 2266 | 82.3 | $0.367 | 4.1 min | #18 |
| GPT-6 Astra (max) | 2185 | 2145 to 2223 | 82.3 | $1.15 | 9.2 min | #21 |
| GPT-6 Astra (medium) | 2003 | 1968 to 2035 | 80.7 | $0.262 | 1.5 min | #32 |
Cost is one script at the API list price of the route used, uncached and without judging. Time is the average for one accepted script, retries included.
Computed Findings Across Thinking Levels
Each finding is computed from the configurations above. Where two 95% ranges overlap, the board cannot order those two configurations.
- Best Configurations Head to Head. GPT-5.6 Sol (low) at Elo 2218 (range 2181 to 2251) and GPT-6 Astra (max) at Elo 2185 (range 2145 to 2223) have overlapping 95% ranges, so the board cannot separate the two best configurations.
- Lowest Thinking Levels Head to Head. At their lowest thinking levels, GPT-5.6 Sol (low) at Elo 2218 and $0.113 per script leads GPT-6 Astra (medium) at Elo 2003 and $0.262 per script by 215 Elo. Their 95% ranges do not overlap.
- Cheapest GPT-5.6 Sol Configuration Above GPT-6 Astra's Best. The cheapest GPT-5.6 Sol configuration with a higher Writing Elo than GPT-6 Astra (max), at Elo 2185 and $1.15 per script, is GPT-5.6 Sol (low), at Elo 2218 and $0.113 per script. Their 95% ranges overlap, so the board cannot separate them. It costs less than the configuration it beats.
- Lowest to Highest Thinking Level. GPT-5.6 Sol moves -5 Elo from low to ultra, at about 3.2 times the cost per script. GPT-6 Astra moves +182 Elo from medium to max, at about 4.4 times the cost per script.
Two Chosen Configurations Compared
Pick one configuration of each model to compare their metric scores, consistency across drafts and per-script scores. Each menu starts on the model's best configuration.
Scores by Metric
GPT-5.6 Sol (low) scores higher on five of the nine metrics and GPT-6 Astra (max) on four. The weight beside each metric is its share of the Writing Score.
Consistency Across Drafts
| GPT-5.6 Sol (low) | GPT-6 Astra (max) | |
|---|---|---|
| Score spread across drafts (±) | 1.27 | 1.87 |
| Lowest script score | 79.8 Script 4 | 80.6 Script 1 |
| Highest script score | 84.2 Script 9 | 85.8 Script 9 |
| Range across scripts | 4.4 | 5.2 |
Spread is how much the Writing Score moves between a configuration's drafts, and lower is steadier. Green marks the better number in each row.
Scores by Script
Writing Score on each reference script. Gap is GPT-5.6 Sol (low) minus GPT-6 Astra (max), and the higher score is in bold.
| Script | GPT-5.6 Sol (low) | GPT-6 Astra (max) | Gap |
|---|---|---|---|
| Script 1 opinion / warning explainer | 82.1 | 80.6 | +1.5 |
| Script 2 news-analysis / skeptical explainer | 82.5 | 82.0 | +0.5 |
| Script 3 personal roadmap / opinion | 82.3 | 84.3 | -2.0 |
| Script 4 short explainer | 79.8 | 82.7 | -2.9 |
| Script 5 news-analysis / opinion explainer | 83.2 | 81.3 | +1.9 |
| Script 6 founder announcement / personal origin story | 82.8 | 83.6 | -0.8 |
| Script 7 personal technical walkthrough / agentic workflow case study | 82.0 | 81.6 | +0.4 |
| Script 8 product release announcement / personal observation | 82.6 | 80.6 | +2.0 |
| Script 9 career guide / hiring analysis | 84.2 | 85.8 | -1.6 |
| Script 10 engineering process walkthrough / presentation adaptation | 82.0 | 80.7 | +1.3 |
Scorecards and Thinking Pages
Each configuration name in the table above opens its own scorecard.
Compare With Another Model
Pick a second model. A pair with its own written match-up opens that page, and any other pair opens in the comparison tool.
Nearby Match-Ups by Best Writing Elo
Every configuration wrote five drafts for each of the same 10 briefs, from the same prompt. Generation was closed-book, with the prompt as the only source and no web search or other tools. Each judge scores every draft once: Claude Opus 5.5 at medium effort through the Claude Code subscription, GPT-6.1 Sol at medium reasoning through the Codex CLI, GLM 5.3 Flash through OpenRouter, and Jev 1.13, a typed judge from TypeSafe that answers each metric as a multiple-choice question. The channel's finished script is their reference for the voice, and it gets no score on the board itself.
← All comparisons