Rewrite · 5 checks · max 400 tokens · benched 2026-09-16

Claude Opus 5 vs Claude Sonnet 5 vs Claude Sonnet 4.6 vs Claude Haiku 4.5 on rewrite a stiff notice in a friendly tone

Anthropic models side by side on "Rewrite a stiff notice in a friendly tone": Claude Haiku 4.5 scores 10/10; Claude Haiku 4.5 is the cheapest answer scoring 8+ at $0.47 per 1,000 runs. Outputs, checks, judge reasons, latency and cost.

The prompt every model received

User

Rewrite the notice below in a warm, friendly, plain-English tone for a company-wide chat message. Keep every fact, including the date and time, in at most 80 words. Do not use the words "hereby" or "pursuant". Output only the rewritten message.

Notice: Staff are hereby informed that, pursuant to the scheduled electrical maintenance, the third-floor kitchen will be unavailable on Friday 3 October between 08:00 and 11:00. Alternative facilities are available on the second floor. Your cooperation is appreciated.

Rubric for the judge: Genuinely friendly tone, all facts kept (date, times, floor, alternative kitchen), under 80 words, no banned words.

Side by side

Every cell is one OpenRouter call at temperature 0 with the prompt's token cap and reasoning effort "low" where the model supports it. Cost is usage × the catalogue rate in models.json. Quality is one judge call to anthropic/claude-haiku-4.5 against the prompt's rubric, cached per prompt version.

Claude Opus 5

anthropic/claude-opus-5

9/10

Latency: 1.5s
Cost: $0.00460
Per 1,000: $4.60

185 in · 147 out · 66 words · checks 5/5

max_words
contains
contains
not_contains
not_contains

Hey everyone! 👋

Quick heads-up: the third-floor kitchen will be closed for scheduled electrical maintenance on Friday 3 October, from 08:00 to 11:00.

In the meantime, feel free to use the kitchen on the second floor — it's all yours for coffee runs, lunches, and everything in between.

Thanks so much for your patience and flexibility that morning. Everything should be back to normal by 11:00!

Judge: Warm, friendly tone with all facts preserved (date, time, floors, alternative location), 67 words, no banned words, though the final sentence adds minor interpretation beyond the original notice.

Claude Sonnet 5

anthropic/claude-sonnet-5

10/10

Latency: 1.1s
Cost: $0.00152
Per 1,000: $1.52

185 in · 115 out · 45 words · checks 3/5

max_words
contains
contains
not_contains
not_contains

Hi everyone! 👋 Just a heads-up: the 3rd-floor kitchen will be closed this Friday, October 3rd, from 8:00–11:00 AM for some electrical maintenance work. Need a coffee or snack break during that time? Head to the 2nd-floor kitchen instead. Thanks so much for your understanding!

Judge: Perfectly rewrites the notice in warm, friendly plain English, includes all facts (date, time, floors, reason), stays well under 80 words (59 words), avoids banned words, and uses appropriate emoji for tone.

Claude Sonnet 4.6

anthropic/claude-sonnet-4.6

10/10

Latency: 988ms
Cost: $0.00158
Per 1,000: $1.57

125 in · 80 out · 51 words · checks 4/5

max_words
contains
contains
not_contains
not_contains

Hey everyone! Just a heads-up — the third-floor kitchen will be out of action this Friday, 3 October, from 8:00 to 11:00 AM due to scheduled electrical maintenance. Not to worry though, the second-floor kitchen is open and ready to use in the meantime. Thanks so much for your understanding! ☕

Judge: Perfectly rewrites the notice in warm, friendly plain English; includes all facts (date, time, floors, reason); 58 words (well under 80); avoids banned words; uses appropriate emoji for tone.

Claude Haiku 4.5

anthropic/claude-haiku-4.5

10/10

top judge scorecheapest scoring 8+

Latency: 709ms
Cost: $0.00047
Per 1,000: $0.47

125 in · 68 out · 44 words · checks 4/5

max_words
contains
contains
not_contains
not_contains

Hey team! Just a heads up—the third-floor kitchen will be closed this Friday, October 3rd from 8:00 AM to 11:00 AM for electrical maintenance. No worries though! Feel free to use the kitchen on the second floor instead. Thanks for your patience and understanding!

Judge: Rewrites notice in warm, friendly plain English; includes all facts (date, time, floors, reason); 48 words (well under 80); avoids banned words; perfectly matches rubric requirements.

Frequently asked

What does this prompt test?

Rewrite: Genuinely friendly tone, all facts kept (date, times, floor, alternative kitchen), under 80 words, no banned words. The deterministic checks are max_words, contains, contains, not_contains, not_contains.

Which model should I pick for this task?

If the judge's bar of 8/10 is good enough for you, Claude Haiku 4.5 at $0.47 per 1,000 runs. If you need the top score, Claude Haiku 4.5 at $0.47 per 1,000 runs.

Same prompt, other families

Run it yourself

Re-run this prompt on Anthropic with your own OpenRouter key, or tweak the wording and see what changes.

Side by side

Claude Opus 5

Claude Sonnet 5

Claude Sonnet 4.6

Claude Haiku 4.5

Frequently asked

Same prompt, other families

Stop overpaying for AI tools