Model Comparison
Claude Haiku 5.5 vs DeepSeek V4 Pro 0813
Too close to call: the best configurations, Claude Haiku 5.5 (max effort) and DeepSeek V4 Pro 0813 (max), are 59 Elo apart, and their 95% ranges overlap. Among 78 models, each ranked by its best configuration, Claude Haiku 5.5 is #16 and DeepSeek V4 Pro 0813 is #21. Cost per script: $0.103 for Claude Haiku 5.5 (max effort) and $0.077 for DeepSeek V4 Pro 0813 (max).
Last updated , from board data generated at 01:13 UTC.
Picks and Verdict
PickClaude Haiku 5.5
Claude Haiku 5.5 (xhigh)
Pick Claude Haiku 5.5 (xhigh) for everyday scripts, since it scores 80.0 on Writing Score against DeepSeek's 80.1 for $0.029 a script and under four minutes.
PickDeepSeek V4 Pro 0813
DeepSeek V4 Pro 0813 (max)
Pick DeepSeek V4 Pro 0813 (max) when the opening matters most, since its 80.9 on Hook Strength leads every Haiku setting for $0.077 a script.
Claude Haiku 5.5 (xhigh) lands a tenth of a point from DeepSeek V4 Pro 0813 (max) on Writing Score, 80.0 against 80.1, for $0.029 a script instead of $0.077. They're 23 Elo apart with overlapping 95% ranges, so the board can't separate them, and Haiku finishes in under four minutes against 8.9. Going up to Claude Haiku 5.5 (max effort) gets a 59 Elo lead on DeepSeek, the #16 model of 78 against the #21 model by best setting, but those ranges overlap too, and it costs $0.103 and more than twelve minutes. That's a lot of waiting for a lead the board can't confirm. Both models need the thinking: Haiku gains 505 Elo from low to max effort, and DeepSeek gains 600 from high to max. The scripts do differ. DeepSeek writes the stronger hook, about three points ahead of Haiku at max effort and more at xhigh, and follows YouTube structure about two points better. Haiku brings more substance and a Slop Score about four and a half points higher. I'd run Haiku at xhigh and rewrite the hook myself.
Disclosure: Claude Haiku 5.5 comes from the same lab as one of the four judges, Claude Opus 5.5. The other three sit outside its lab, so it is never 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 Claude Haiku 5.5 and DeepSeek V4 Pro 0813, 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 | Reasoning Tokens | Configuration Rank |
|---|---|---|---|---|---|---|---|
| Claude Haiku 5.5 (max effort) | 2009 | 1972 to 2044 | 80.8 | $0.103 | 12.6 min | N/A | #31 |
| DeepSeek V4 Pro 0813 (max) | 1950 | 1906 to 1989 | 80.1 | $0.077 | 8.9 min | 31,573 | #40 |
| Claude Haiku 5.5 (xhigh) | 1927 | 1883 to 1965 | 80.0 | $0.029 | 3.6 min | N/A | #41 |
| Claude Haiku 5.5 (high) | 1723 | 1679 to 1760 | 78.0 | $0.012 | 1.5 min | N/A | #53 |
| Claude Haiku 5.5 (medium) | 1620 | 1572 to 1670 | 76.7 | $0.0067 | 0.8 min | N/A | #62 |
| Claude Haiku 5.5 (low) | 1504 | 1461 to 1550 | 75.2 | $0.0045 | 0.5 min | N/A | #74 |
| DeepSeek V4 Pro 0813 (high) | 1350 | 1272 to 1428 | 72.3 | $0.031 | 1.3 min | 8,879 | #81 |
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. N/A under Reasoning Tokens means the board has no separate reasoning count for that route.
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. Claude Haiku 5.5 (max effort) at Elo 2009 (range 1972 to 2044) and DeepSeek V4 Pro 0813 (max) at Elo 1950 (range 1906 to 1989) have overlapping 95% ranges, so the board cannot separate the two best configurations.
- Lowest Thinking Levels Head to Head. At their lowest thinking levels, Claude Haiku 5.5 (low) at Elo 1504 and $0.0045 per script leads DeepSeek V4 Pro 0813 (high) at Elo 1350 and $0.031 per script by 154 Elo. Their 95% ranges do not overlap.
- Same Level Name, Side by Side. Claude Haiku 5.5 has the higher Writing Elo at every level name both models use: high and max. The 95% ranges do not overlap at high. Level names come from each provider, so the same name does not promise the same reasoning budget.
- Cheapest Configuration Overlapping the Top One. DeepSeek V4 Pro 0813 (max), at $0.077 per script, is the cheapest configuration whose 95% range overlaps that of Claude Haiku 5.5 (max effort), at $0.103 per script: 1906 to 1989 against 1972 to 2044, so the board cannot separate them. Claude Haiku 5.5 (max effort) costs about 1.3 times as much per script.
- Cheapest Claude Haiku 5.5 Configuration Above DeepSeek V4 Pro 0813's Best. The cheapest Claude Haiku 5.5 configuration with a higher Writing Elo than DeepSeek V4 Pro 0813 (max), at Elo 1950 and $0.077 per script, is Claude Haiku 5.5 (max effort), at Elo 2009 and $0.103 per script. Their 95% ranges overlap, so the board cannot separate them. It costs about 1.3 times as much per script.
- Lowest to Highest Thinking Level. Claude Haiku 5.5 moves +505 Elo from low to max effort, at about 23 times the cost per script. DeepSeek V4 Pro 0813 moves +600 Elo from high to max, at about 2.5 times the cost per script.
Writing Elo at Each Shared Level Name
| Level | Claude Haiku 5.5 | DeepSeek V4 Pro 0813 | Gap | 95% Ranges |
|---|---|---|---|---|
| high | 1723 | 1350 | +373 | Separate |
| max | 2009 | 1950 | +59 | Overlap |
Gap is Claude Haiku 5.5 minus DeepSeek V4 Pro 0813 at the same level name, using each model's highest-rated configuration at that level.
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
Claude Haiku 5.5 (max effort) scores higher on six of the nine metrics and DeepSeek V4 Pro 0813 (max) on three. Weighted by each metric's share of the Writing Score, Substance, Accuracy & Value does the most to put Claude Haiku 5.5 (max effort) ahead, with a 3.9-point lead on that metric. The weight beside each metric is its share of the Writing Score.
Consistency Across Drafts
| Claude Haiku 5.5 (max effort) | DeepSeek V4 Pro 0813 (max) | |
|---|---|---|
| Score spread across drafts (±) | 1.55 | 1.48 |
| Lowest script score | 78.2 Script 10 | 78.2 Script 10 |
| Highest script score | 82.6 Script 4 | 81.5 Script 6 |
| Range across scripts | 4.4 | 3.3 |
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 Claude Haiku 5.5 (max effort) minus DeepSeek V4 Pro 0813 (max), and the higher score is in bold.
| Script | Claude Haiku 5.5 (max effort) | DeepSeek V4 Pro 0813 (max) | Gap |
|---|---|---|---|
| Script 1 opinion / warning explainer | 80.3 | 79.5 | +0.8 |
| Script 2 news-analysis / skeptical explainer | 80.7 | 80.6 | +0.1 |
| Script 3 personal roadmap / opinion | 80.8 | 80.8 | 0.0 |
| Script 4 short explainer | 82.6 | 78.5 | +4.1 |
| Script 5 news-analysis / opinion explainer | 81.7 | 80.9 | +0.8 |
| Script 6 founder announcement / personal origin story | 81.1 | 81.5 | -0.4 |
| Script 7 personal technical walkthrough / agentic workflow case study | 80.0 | 81.1 | -1.1 |
| Script 8 product release announcement / personal observation | 81.3 | 80.2 | +1.1 |
| Script 9 career guide / hiring analysis | 81.5 | 79.9 | +1.6 |
| Script 10 engineering process walkthrough / presentation adaptation | 78.2 | 78.2 | 0.0 |
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