Whet Benchmark

How much can an LLM sharpen a prompt without destroying its intent?

Meta-prompt-following under pressure to preserve purpose — before/after score delta for each model. Results are measured using Whet's internal prompts and criteria: they don't constitute an academic benchmark or official comparison. Useful as a reference for real behavior; not as a definitive verdict.

  • A gap MMLU, HumanEval and HLE don't cover.
  • Open methodology · open data · open code.
  • Before-after delta. No stunts, no cherry-picking.
Two benchmarks in Whet

A technical axis (prompt sharpening) and a political axis (ideological positioning). Same open methodology, two distinct cuts.

Models tested
14
Prompts evaluated
74
44 concepts · cumulative
Average delta
+45.1
Last run
May 08, 2026
corpus v3

Leaderboard

sorted by delta ↓
#modellabbeforeafterΔlatencyprompts
1
Jamba Large 1.7 (AI21)
jamba-large-1.7
AI21 Labs (Israel)trial
Arquitetura Mamba-Transformer híbrida — única não-transformer pura do benchmark
42.694.0+51.44.5s74
2
Claude Sonnet (via CLI)
claude-sonnet-4-6 (via CLI)
Anthropic (EUA)paid
Claude Sonnet via CLI — versão mais rápida e leve que Opus
42.693.6+51.020.8s74
3
Claude Opus (via CLI)
claude-opus-4-7 (via CLI)
Anthropic (EUA)paid
Claude Opus via CLI subscription — sem API key separada
42.693.1+50.414.4s74
4
Grok 4.20 Reasoning (xAI)
grok-4.20-reasoning
xAI (USA)data-sharing
Self-described less-restrictive alignment — tests meta-prompt-following under lower resistance to reformulating
42.691.0+48.318.7s74
5
GPT-5.4 (OpenAI)
gpt-5.4
OpenAI (EUA)paid
Flagship atual da OpenAI (lançado março/2026) — testa se a nova geração quebra o padrão conservador do gpt-4o-mini
42.689.8+47.15.0s74
6
GPT-5.5 (OpenAI)
gpt-5.5
OpenAI (EUA)paid
Iteração flagship reasoning sucessora do gpt-5.4 — mede se o salto pra reasoning por padrão na linhagem 5.5 mantém o ganho de meta-prompt-following
42.689.6+47.07.9s74
7
DeepSeek R1
deepseek-reasoner
DeepSeek (China)free
Versão reasoner do DeepSeek com chain-of-thought explícito
42.688.9+46.339.9s74
8
Mistral Small
mistral-small-latest
Mistral AI (França)free
Modelo europeu open-weight focado em eficiência
42.688.9+46.22.4s74
9
Llama 3.3 70B (Groq)
llama-3.3-70b-versatile
Meta (EUA) via Groqfree
Llama open-source com inferência em hardware LPU da Groq
42.687.9+45.24.0s74
10
DeepSeek V3
deepseek-chat
DeepSeek (China)free
Modelo chinês de alta performance treinado com foco em código e raciocínio
42.687.0+44.37.9s74
11
Command A (Cohere)
command-a-03-2025
Cohere (Canadá)trial
Filosofia de alinhamento focada em RAG/grounded generation/tool use — testa meta-prompt-following sob viés distinto
42.683.9+41.39.5s74
12
Gemini 2.5 Flash
gemini-2.5-flash
Google (EUA)free
Modelo multimodal de referência do Google
42.681.7+39.08.6s74
13
GPT-5 nano (OpenAI)
gpt-5-nano
OpenAI (EUA)paid
Tier nano da geração GPT-5 — barato e recente, contraste direto ao gpt-4o-mini legado
42.680.4+37.85.8s74
14
GPT-4o mini (OpenAI)
gpt-4o-mini
OpenAI (EUA)paid
Flagship cost-efficient da OpenAI — representa a filosofia RLHF pioneira do campo
42.678.5+35.93.8s74

Per-prompt breakdown

Each provider appears as a column. Click a provider to open the expanded list with the result for each corpus prompt — useful for understanding where each model sharpens better or worse.

Jamba Large 1.7 (AI21)74 prompts · avg Δ +51.4
promptlangbeforeafterΔ
legal-advisor-en?0100+100
rigid-code-agent-en?1100+99
rigid-code-agent-pt?197+96
agente-codigo-rigido?197+96
legal-advisor-pt?093+93
corporate-counsel-pt?093+93
clinical-nurse-intake-en?093+93
personal-investment-advisor-es?093+93
consultor-juridico?090+90
medical-assistant-pt?087+87
legal-advisor-es?14100+86
medical-assistant-en?086+86
disguised-repetitions-pt?1397+84
repeticoes-disfarcadas?1397+84
assistente-medico?080+80
UX lead pre-IPO Avenida Pay (es)es24100+76
rigid-code-agent-es?2297+75
disguised-repetitions-en?2397+74
marketing-strategist-pt?28100+72
estrategista-marketing?28100+72
ux-research-assistant-en?2597+72
social-media-agency-pt?2090+70
medical-assistant-es?2493+69
disguised-repetitions-es?2897+69
Mediación familiar / uniones de hecho (es)es2993+64
Satellite ground-station CI reviewer (en)en38100+62
Assistente de leilão judicial de imóveis (pt)pt1676+60
Cybersec bootcamp instructional designer (en)en41100+59
marketing-strategist-en?43100+57
PMO de hortas hidropônicas em escolas (pt)pt44100+56
Coordinador last-mile en Lima (es)es4297+55
Pediatric dental clinic intake (en)en3893+55
newsroom-writer-es?4397+54
logistics-optimizer-en?47100+53
financial-analyst-en?3790+53
Contadora tributária para usinas de etanol (pt)pt4093+53
tax-consultant-en?48100+52
Tangier waterfront pier competition (en)en3486+52
education-tutor-es?49100+51
tax-consultant-es?51100+49
financial-analyst-pt?3180+49
education-tutor-pt?4997+48
tone-domain-mismatch-es?4593+48
logistics-optimizer-pt?55100+45
education-tutor-en?5397+44
tax-consultant-pt?5193+42
task-style-no-roleplay-pt?5290+38
nutripal-weight-loss-en?4886+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
logistics-optimizer-es?63100+37
marketing-strategist-es?64100+36
task-style-no-roleplay-en?4883+35
mental-health-wellness-es?5893+35
Cierre y remediación de pasivos mineros (es)es5590+35
financial-analyst-es?5690+34
task-style-no-roleplay-es?6597+32
virtual-pharmacist-pt?6593+28
cooperative-agronomist-es?6693+27
contradiction-es?6186+25
senior-structural-engineer-pt?77100+23
streaming-script-editor-en?6486+22
unresolved-reference-es?79100+21
contradiction-en?6179+18
threat-framing-en?83100+17
unresolved-reference-pt?6986+17
threat-framing-es?7993+14
unresolved-reference-en?7690+14
role-inflation-pt?88100+12
contradiction-pt?6172+11
tone-domain-mismatch-pt?93100+7
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
threat-framing-pt?9093+3
role-inflation-es?97100+3
Claude Sonnet (via CLI)74 prompts · avg Δ +51.0
promptlangbeforeafterΔ
medical-assistant-pt?094+94
legal-advisor-en?093+93
personal-investment-advisor-es?093+93
consultor-juridico?090+90
agente-codigo-rigido?190+89
disguised-repetitions-pt?13100+87
repeticoes-disfarcadas?13100+87
legal-advisor-pt?086+86
rigid-code-agent-en?187+86
corporate-counsel-pt?086+86
medical-assistant-en?083+83
assistente-medico?083+83
clinical-nurse-intake-en?083+83
rigid-code-agent-pt?181+80
social-media-agency-pt?20100+80
disguised-repetitions-en?23100+77
ux-research-assistant-en?25100+75
disguised-repetitions-es?28100+72
Mediación familiar / uniones de hecho (es)es29100+71
legal-advisor-es?1479+65
rigid-code-agent-es?2287+65
estrategista-marketing?2893+65
medical-assistant-es?2486+62
UX lead pre-IPO Avenida Pay (es)es2486+62
Assistente de leilão judicial de imóveis (pt)pt1675+59
Cybersec bootcamp instructional designer (en)en41100+59
Tangier waterfront pier competition (en)en3493+59
marketing-strategist-pt?2886+58
Coordinador last-mile en Lima (es)es42100+58
newsroom-writer-es?43100+57
marketing-strategist-en?4397+54
logistics-optimizer-en?47100+53
PMO de hortas hidropônicas em escolas (pt)pt4497+53
Contadora tributária para usinas de etanol (pt)pt4093+53
financial-analyst-pt?3183+52
task-style-no-roleplay-en?48100+52
tax-consultant-en?48100+52
nutripal-weight-loss-en?48100+52
Satellite ground-station CI reviewer (en)en3890+52
Pediatric dental clinic intake (en)en3890+52
education-tutor-pt?49100+51
education-tutor-es?49100+51
tax-consultant-es?51100+49
tone-domain-mismatch-es?4593+48
education-tutor-en?53100+47
financial-analyst-en?3783+46
tax-consultant-pt?5193+42
mental-health-wellness-es?58100+42
task-style-no-roleplay-pt?5293+41
contradiction-en?61100+39
logistics-optimizer-pt?5593+38
Cierre y remediación de pasivos mineros (es)es5593+38
financial-analyst-es?5693+37
logistics-optimizer-es?63100+37
contradiction-pt?6197+36
contradiction-es?6197+36
marketing-strategist-es?6497+33
virtual-pharmacist-pt?6597+32
unresolved-reference-pt?69100+31
Integridade de risers em águas ultraprofundas (pt)pt6293+31
task-style-no-roleplay-es?6593+28
cooperative-agronomist-es?6693+27
streaming-script-editor-en?6490+26
unresolved-reference-en?76100+24
unresolved-reference-es?79100+21
senior-structural-engineer-pt?7797+20
threat-framing-en?83100+17
threat-framing-es?7993+14
threat-framing-pt?90100+10
tone-domain-mismatch-en?93100+7
role-inflation-es?97100+3
role-inflation-pt?8887+-1
role-inflation-en?9490+-4
tone-domain-mismatch-pt?9379+-14
Claude Opus (via CLI)74 prompts · avg Δ +50.4
promptlangbeforeafterΔ
consultor-juridico?0100+100
legal-advisor-en?0100+100
legal-advisor-pt?0100+100
assistente-medico?093+93
medical-assistant-en?093+93
medical-assistant-pt?090+90
corporate-counsel-pt?090+90
clinical-nurse-intake-en?090+90
rigid-code-agent-pt?190+89
repeticoes-disfarcadas?13100+87
agente-codigo-rigido?187+86
rigid-code-agent-en?187+86
personal-investment-advisor-es?086+86
disguised-repetitions-pt?1397+84
social-media-agency-pt?20100+80
legal-advisor-es?1493+79
disguised-repetitions-en?23100+77
medical-assistant-es?24100+76
disguised-repetitions-es?2897+69
ux-research-assistant-en?2593+68
Assistente de leilão judicial de imóveis (pt)pt1683+67
estrategista-marketing?2893+65
rigid-code-agent-es?2284+62
Cybersec bootcamp instructional designer (en)en41100+59
UX lead pre-IPO Avenida Pay (es)es2483+59
marketing-strategist-pt?2886+58
Mediación familiar / uniones de hecho (es)es2986+57
financial-analyst-en?3793+56
task-style-no-roleplay-en?48100+52
tax-consultant-en?48100+52
Satellite ground-station CI reviewer (en)en3890+52
Tangier waterfront pier competition (en)en3486+52
education-tutor-pt?49100+51
education-tutor-es?49100+51
Coordinador last-mile en Lima (es)es4293+51
marketing-strategist-en?4393+50
newsroom-writer-es?4393+50
Contadora tributária para usinas de etanol (pt)pt4090+50
tax-consultant-es?51100+49
PMO de hortas hidropônicas em escolas (pt)pt4493+49
tone-domain-mismatch-es?4593+48
task-style-no-roleplay-pt?52100+48
education-tutor-en?53100+47
logistics-optimizer-en?4793+46
logistics-optimizer-pt?55100+45
nutripal-weight-loss-en?4893+45
tax-consultant-pt?5193+42
mental-health-wellness-es?58100+42
financial-analyst-pt?3172+41
contradiction-pt?61100+39
contradiction-es?61100+39
Integridade de risers em águas ultraprofundas (pt)pt62100+38
Pediatric dental clinic intake (en)en3875+37
contradiction-en?6193+32
unresolved-reference-pt?69100+31
Cierre y remediación de pasivos mineros (es)es5586+31
task-style-no-roleplay-es?6593+28
financial-analyst-es?5683+27
logistics-optimizer-es?6390+27
cooperative-agronomist-es?6693+27
unresolved-reference-en?76100+24
senior-structural-engineer-pt?77100+23
streaming-script-editor-en?6487+23
unresolved-reference-es?79100+21
virtual-pharmacist-pt?6586+21
marketing-strategist-es?6479+15
threat-framing-es?7993+14
threat-framing-en?8393+10
role-inflation-en?94100+6
role-inflation-pt?8893+5
threat-framing-pt?9093+3
tone-domain-mismatch-pt?9393+0
tone-domain-mismatch-en?9390+-3
role-inflation-es?9793+-4
Grok 4.20 Reasoning (xAI)74 prompts · avg Δ +48.3
promptlangbeforeafterΔ
legal-advisor-en?0100+100
legal-advisor-pt?093+93
medical-assistant-en?093+93
consultor-juridico?093+93
disguised-repetitions-pt?13100+87
repeticoes-disfarcadas?13100+87
legal-advisor-es?14100+86
assistente-medico?086+86
rigid-code-agent-pt?184+83
agente-codigo-rigido?184+83
medical-assistant-pt?079+79
corporate-counsel-pt?079+79
clinical-nurse-intake-en?079+79
rigid-code-agent-en?176+75
ux-research-assistant-en?25100+75
disguised-repetitions-en?2397+74
social-media-agency-pt?2093+73
disguised-repetitions-es?28100+72
personal-investment-advisor-es?072+72
rigid-code-agent-es?2290+68
Tangier waterfront pier competition (en)en34100+66
medical-assistant-es?2486+62
estrategista-marketing?2890+62
UX lead pre-IPO Avenida Pay (es)es2486+62
financial-analyst-en?3797+60
Cybersec bootcamp instructional designer (en)en41100+59
marketing-strategist-en?43100+57
tone-domain-mismatch-es?45100+55
Satellite ground-station CI reviewer (en)en3893+55
logistics-optimizer-en?47100+53
tax-consultant-en?48100+52
task-style-no-roleplay-en?48100+52
nutripal-weight-loss-en?48100+52
education-tutor-pt?49100+51
Coordinador last-mile en Lima (es)es4293+51
tax-consultant-es?51100+49
PMO de hortas hidropônicas em escolas (pt)pt4493+49
marketing-strategist-pt?2876+48
financial-analyst-pt?3179+48
task-style-no-roleplay-pt?52100+48
education-tutor-en?53100+47
newsroom-writer-es?4390+47
Pediatric dental clinic intake (en)en3883+45
tax-consultant-pt?5193+42
mental-health-wellness-es?58100+42
contradiction-pt?61100+39
logistics-optimizer-pt?5593+38
Assistente de leilão judicial de imóveis (pt)pt1654+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
education-tutor-es?4986+37
Mediación familiar / uniones de hecho (es)es2966+37
marketing-strategist-es?64100+36
contradiction-en?6193+32
contradiction-es?6193+32
Contadora tributária para usinas de etanol (pt)pt4072+32
Cierre y remediación de pasivos mineros (es)es5586+31
logistics-optimizer-es?6393+30
financial-analyst-es?5686+30
virtual-pharmacist-pt?6593+28
cooperative-agronomist-es?6693+27
streaming-script-editor-en?6490+26
unresolved-reference-es?79100+21
threat-framing-en?83100+17
unresolved-reference-pt?6986+17
unresolved-reference-en?7693+17
task-style-no-roleplay-es?6579+14
senior-structural-engineer-pt?7790+13
role-inflation-pt?88100+12
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
role-inflation-es?97100+3
threat-framing-es?7979+0
threat-framing-pt?9086+-4
tone-domain-mismatch-pt?9365+-28
GPT-5.4 (OpenAI)74 prompts · avg Δ +47.1
promptlangbeforeafterΔ
medical-assistant-en?090+90
clinical-nurse-intake-en?090+90
medical-assistant-pt?086+86
assistente-medico?086+86
agente-codigo-rigido?187+86
legal-advisor-en?084+84
legal-advisor-pt?083+83
rigid-code-agent-en?184+83
disguised-repetitions-pt?1394+81
repeticoes-disfarcadas?1394+81
corporate-counsel-pt?080+80
personal-investment-advisor-es?080+80
rigid-code-agent-pt?180+79
consultor-juridico?079+79
disguised-repetitions-en?23100+77
ux-research-assistant-en?25100+75
Assistente de leilão judicial de imóveis (pt)pt1690+74
legal-advisor-es?1486+72
rigid-code-agent-es?2293+71
social-media-agency-pt?2090+70
medical-assistant-es?2493+69
disguised-repetitions-es?2897+69
UX lead pre-IPO Avenida Pay (es)es2493+69
Mediación familiar / uniones de hecho (es)es2986+57
financial-analyst-pt?3186+55
estrategista-marketing?2883+55
financial-analyst-en?3790+53
tax-consultant-en?48100+52
Satellite ground-station CI reviewer (en)en3890+52
education-tutor-pt?49100+51
education-tutor-es?49100+51
Coordinador last-mile en Lima (es)es4293+51
newsroom-writer-es?4393+50
task-style-no-roleplay-en?4897+49
tone-domain-mismatch-es?4593+48
task-style-no-roleplay-pt?52100+48
education-tutor-en?53100+47
PMO de hortas hidropônicas em escolas (pt)pt4490+46
Cybersec bootcamp instructional designer (en)en4187+46
logistics-optimizer-en?4790+43
tax-consultant-pt?5193+42
tax-consultant-es?5193+42
mental-health-wellness-es?58100+42
Tangier waterfront pier competition (en)en3476+42
marketing-strategist-pt?2869+41
contradiction-en?61100+39
contradiction-es?61100+39
nutripal-weight-loss-en?4886+38
Cierre y remediación de pasivos mineros (es)es5593+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
financial-analyst-es?5693+37
task-style-no-roleplay-es?65100+35
marketing-strategist-en?4376+33
Contadora tributária para usinas de etanol (pt)pt4073+33
contradiction-pt?6193+32
logistics-optimizer-pt?5586+31
logistics-optimizer-es?6393+30
Pediatric dental clinic intake (en)en3868+30
cooperative-agronomist-es?6693+27
threat-framing-es?79100+21
marketing-strategist-es?6483+19
unresolved-reference-es?7997+18
virtual-pharmacist-pt?6583+18
unresolved-reference-pt?6986+17
threat-framing-en?8397+14
senior-structural-engineer-pt?7790+13
role-inflation-pt?88100+12
threat-framing-pt?9097+7
unresolved-reference-en?7683+7
role-inflation-en?94100+6
streaming-script-editor-en?6470+6
role-inflation-es?97100+3
tone-domain-mismatch-en?9383+-10
tone-domain-mismatch-pt?9372+-21
GPT-5.5 (OpenAI)74 prompts · avg Δ +47.0
promptlangbeforeafterΔ
legal-advisor-en?090+90
medical-assistant-pt?090+90
consultor-juridico?090+90
assistente-medico?090+90
corporate-counsel-pt?090+90
legal-advisor-pt?087+87
medical-assistant-en?087+87
personal-investment-advisor-es?087+87
rigid-code-agent-pt?187+86
rigid-code-agent-en?187+86
disguised-repetitions-pt?1397+84
repeticoes-disfarcadas?1397+84
agente-codigo-rigido?184+83
clinical-nurse-intake-en?079+79
disguised-repetitions-en?23100+77
social-media-agency-pt?2090+70
UX lead pre-IPO Avenida Pay (es)es2493+69
legal-advisor-es?1482+68
rigid-code-agent-es?2290+68
ux-research-assistant-en?2590+65
estrategista-marketing?2890+62
Assistente de leilão judicial de imóveis (pt)pt1677+61
medical-assistant-es?2483+59
Tangier waterfront pier competition (en)en3493+59
disguised-repetitions-es?2884+56
Satellite ground-station CI reviewer (en)en3893+55
tax-consultant-en?48100+52
task-style-no-roleplay-en?48100+52
education-tutor-pt?49100+51
education-tutor-es?49100+51
financial-analyst-pt?3182+51
newsroom-writer-es?4393+50
tax-consultant-es?51100+49
task-style-no-roleplay-pt?52100+48
education-tutor-en?53100+47
financial-analyst-en?3782+45
logistics-optimizer-en?4790+43
Contadora tributária para usinas de etanol (pt)pt4083+43
tax-consultant-pt?5193+42
PMO de hortas hidropônicas em escolas (pt)pt4486+42
Mediación familiar / uniones de hecho (es)es2971+42
tone-domain-mismatch-es?4586+41
marketing-strategist-en?4382+39
contradiction-en?61100+39
nutripal-weight-loss-en?4886+38
Coordinador last-mile en Lima (es)es4280+38
marketing-strategist-pt?2865+37
contradiction-es?6197+36
task-style-no-roleplay-es?65100+35
Cierre y remediación de pasivos mineros (es)es5590+35
Integridade de risers em águas ultraprofundas (pt)pt6297+35
cooperative-agronomist-es?66100+34
Pediatric dental clinic intake (en)en3872+34
logistics-optimizer-pt?5586+31
unresolved-reference-pt?69100+31
logistics-optimizer-es?6393+30
Cybersec bootcamp instructional designer (en)en4169+28
marketing-strategist-es?6490+26
financial-analyst-es?5682+26
contradiction-pt?6187+26
senior-structural-engineer-pt?77100+23
unresolved-reference-es?79100+21
streaming-script-editor-en?6484+20
threat-framing-es?7997+18
unresolved-reference-en?7690+14
mental-health-wellness-es?5872+14
role-inflation-pt?88100+12
threat-framing-en?8394+11
virtual-pharmacist-pt?6576+11
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
threat-framing-pt?9094+4
role-inflation-es?97100+3
tone-domain-mismatch-pt?9376+-17
DeepSeek R174 prompts · avg Δ +46.3
promptlangbeforeafterΔ
clinical-nurse-intake-en?0100+100
rigid-code-agent-en?197+96
legal-advisor-en?093+93
legal-advisor-pt?086+86
personal-investment-advisor-es?086+86
agente-codigo-rigido?186+85
repeticoes-disfarcadas?1397+84
consultor-juridico?083+83
disguised-repetitions-pt?1394+81
Assistente de leilão judicial de imóveis (pt)pt1697+81
assistente-medico?079+79
medical-assistant-en?077+77
legal-advisor-es?1490+76
medical-assistant-pt?076+76
corporate-counsel-pt?076+76
UX lead pre-IPO Avenida Pay (es)es24100+76
disguised-repetitions-en?2397+74
rigid-code-agent-pt?174+73
disguised-repetitions-es?28100+72
ux-research-assistant-en?2597+72
Tangier waterfront pier competition (en)en34100+66
estrategista-marketing?2893+65
financial-analyst-pt?3194+63
rigid-code-agent-es?2284+62
medical-assistant-es?2483+59
social-media-agency-pt?2079+59
logistics-optimizer-en?47100+53
task-style-no-roleplay-en?48100+52
Cybersec bootcamp instructional designer (en)en4193+52
financial-analyst-en?3787+50
newsroom-writer-es?4393+50
Mediación familiar / uniones de hecho (es)es2979+50
tax-consultant-en?4897+49
tone-domain-mismatch-es?4593+48
education-tutor-es?4997+48
education-tutor-en?53100+47
PMO de hortas hidropônicas em escolas (pt)pt4490+46
marketing-strategist-en?4387+44
tax-consultant-pt?5193+42
tax-consultant-es?5193+42
mental-health-wellness-es?58100+42
Satellite ground-station CI reviewer (en)en3880+42
task-style-no-roleplay-pt?5293+41
marketing-strategist-pt?2866+38
financial-analyst-es?5694+38
Cierre y remediación de pasivos mineros (es)es5593+38
Pediatric dental clinic intake (en)en3876+38
education-tutor-pt?4986+37
Coordinador last-mile en Lima (es)es4279+37
nutripal-weight-loss-en?4883+35
cooperative-agronomist-es?66100+34
logistics-optimizer-pt?5586+31
Integridade de risers em águas ultraprofundas (pt)pt6293+31
logistics-optimizer-es?6393+30
marketing-strategist-es?6493+29
streaming-script-editor-en?6493+29
contradiction-pt?6186+25
contradiction-en?6186+25
contradiction-es?6186+25
task-style-no-roleplay-es?6587+22
Contadora tributária para usinas de etanol (pt)pt4062+22
unresolved-reference-en?7693+17
threat-framing-es?7993+14
virtual-pharmacist-pt?6579+14
senior-structural-engineer-pt?7790+13
threat-framing-en?8393+10
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
role-inflation-pt?8893+5
unresolved-reference-pt?6972+3
unresolved-reference-es?7979+0
role-inflation-es?9793+-4
threat-framing-pt?9083+-7
tone-domain-mismatch-pt?9379+-14
Mistral Small74 prompts · avg Δ +46.2
promptlangbeforeafterΔ
legal-advisor-en?097+97
medical-assistant-en?093+93
corporate-counsel-pt?093+93
repeticoes-disfarcadas?13100+87
clinical-nurse-intake-en?086+86
rigid-code-agent-en?184+83
legal-advisor-pt?079+79
disguised-repetitions-pt?1390+77
disguised-repetitions-en?23100+77
agente-codigo-rigido?177+76
medical-assistant-pt?076+76
rigid-code-agent-pt?177+76
UX lead pre-IPO Avenida Pay (es)es24100+76
ux-research-assistant-en?25100+75
legal-advisor-es?1486+72
social-media-agency-pt?2090+70
consultor-juridico?069+69
disguised-repetitions-es?2897+69
estrategista-marketing?2894+66
assistente-medico?065+65
personal-investment-advisor-es?065+65
marketing-strategist-pt?2891+63
financial-analyst-pt?3190+59
rigid-code-agent-es?2280+58
Mediación familiar / uniones de hecho (es)es2986+57
medical-assistant-es?2479+55
Satellite ground-station CI reviewer (en)en3893+55
Tangier waterfront pier competition (en)en3486+52
education-tutor-es?49100+51
logistics-optimizer-en?4797+50
newsroom-writer-es?4393+50
PMO de hortas hidropônicas em escolas (pt)pt4493+49
Cybersec bootcamp instructional designer (en)en4190+49
tone-domain-mismatch-es?4593+48
task-style-no-roleplay-pt?52100+48
education-tutor-en?53100+47
financial-analyst-en?3783+46
task-style-no-roleplay-en?4893+45
tax-consultant-en?4893+45
Pediatric dental clinic intake (en)en3883+45
education-tutor-pt?4993+44
Coordinador last-mile en Lima (es)es4286+44
marketing-strategist-en?4386+43
Assistente de leilão judicial de imóveis (pt)pt1657+41
tax-consultant-pt?5190+39
Contadora tributária para usinas de etanol (pt)pt4079+39
nutripal-weight-loss-en?4886+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
logistics-optimizer-es?63100+37
task-style-no-roleplay-es?65100+35
tax-consultant-es?5186+35
streaming-script-editor-en?6497+33
Cierre y remediación de pasivos mineros (es)es5586+31
marketing-strategist-es?6494+30
contradiction-es?6190+29
logistics-optimizer-pt?5583+28
mental-health-wellness-es?5886+28
financial-analyst-es?5683+27
cooperative-agronomist-es?6693+27
contradiction-pt?6183+22
threat-framing-es?79100+21
unresolved-reference-en?7693+17
senior-structural-engineer-pt?7793+16
unresolved-reference-es?7993+14
virtual-pharmacist-pt?6579+14
role-inflation-pt?88100+12
contradiction-en?6173+12
unresolved-reference-pt?6979+10
tone-domain-mismatch-pt?93100+7
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
threat-framing-en?8387+4
role-inflation-es?97100+3
threat-framing-pt?9079+-11
Llama 3.3 70B (Groq)74 prompts · avg Δ +45.2
promptlangbeforeafterΔ
legal-advisor-en?091+91
medical-assistant-en?090+90
repeticoes-disfarcadas?13100+87
rigid-code-agent-en?187+86
disguised-repetitions-pt?1393+80
consultor-juridico?079+79
assistente-medico?079+79
legal-advisor-pt?079+79
medical-assistant-pt?079+79
clinical-nurse-intake-en?076+76
ux-research-assistant-en?25100+75
disguised-repetitions-en?2397+74
social-media-agency-pt?2093+73
UX lead pre-IPO Avenida Pay (es)es2493+69
estrategista-marketing?2893+65
rigid-code-agent-pt?165+64
financial-analyst-en?37100+63
Assistente de leilão judicial de imóveis (pt)pt1679+63
medical-assistant-es?2486+62
marketing-strategist-pt?2890+62
financial-analyst-pt?3193+62
personal-investment-advisor-es?062+62
Satellite ground-station CI reviewer (en)en3897+59
legal-advisor-es?1472+58
disguised-repetitions-es?2886+58
agente-codigo-rigido?158+57
rigid-code-agent-es?2279+57
Mediación familiar / uniones de hecho (es)es2986+57
Cybersec bootcamp instructional designer (en)en4197+56
marketing-strategist-en?4397+54
logistics-optimizer-en?47100+53
task-style-no-roleplay-en?48100+52
tax-consultant-en?48100+52
corporate-counsel-pt?051+51
newsroom-writer-es?4393+50
tax-consultant-es?51100+49
PMO de hortas hidropônicas em escolas (pt)pt4493+49
Contadora tributária para usinas de etanol (pt)pt4086+46
education-tutor-en?5397+44
tax-consultant-pt?5193+42
nutripal-weight-loss-en?4890+42
Tangier waterfront pier competition (en)en3476+42
tone-domain-mismatch-es?4586+41
task-style-no-roleplay-pt?5293+41
Pediatric dental clinic intake (en)en3879+41
contradiction-en?61100+39
education-tutor-pt?4986+37
education-tutor-es?4986+37
task-style-no-roleplay-es?65100+35
Coordinador last-mile en Lima (es)es4276+34
contradiction-es?6193+32
Integridade de risers em águas ultraprofundas (pt)pt6293+31
logistics-optimizer-pt?5583+28
virtual-pharmacist-pt?6593+28
mental-health-wellness-es?5886+28
cooperative-agronomist-es?6693+27
contradiction-pt?6186+25
logistics-optimizer-es?6386+23
unresolved-reference-es?79100+21
financial-analyst-es?5676+20
Cierre y remediación de pasivos mineros (es)es5572+17
marketing-strategist-es?6480+16
streaming-script-editor-en?6479+15
unresolved-reference-en?7690+14
senior-structural-engineer-pt?7790+13
role-inflation-pt?88100+12
threat-framing-en?8390+7
threat-framing-es?7986+7
tone-domain-mismatch-pt?93100+7
tone-domain-mismatch-en?93100+7
role-inflation-en?94100+6
threat-framing-pt?9093+3
role-inflation-es?97100+3
unresolved-reference-pt?6972+3
DeepSeek V374 prompts · avg Δ +44.3
promptlangbeforeafterΔ
legal-advisor-en?097+97
clinical-nurse-intake-en?090+90
agente-codigo-rigido?190+89
disguised-repetitions-pt?13100+87
repeticoes-disfarcadas?13100+87
legal-advisor-pt?086+86
rigid-code-agent-en?187+86
consultor-juridico?083+83
rigid-code-agent-pt?183+82
medical-assistant-pt?079+79
medical-assistant-en?079+79
assistente-medico?079+79
personal-investment-advisor-es?079+79
rigid-code-agent-es?22100+78
disguised-repetitions-en?23100+77
ux-research-assistant-en?25100+75
disguised-repetitions-es?28100+72
corporate-counsel-pt?072+72
legal-advisor-es?1479+65
medical-assistant-es?2486+62
financial-analyst-pt?3193+62
UX lead pre-IPO Avenida Pay (es)es2486+62
Tangier waterfront pier competition (en)en3493+59
Coordinador last-mile en Lima (es)es42100+58
newsroom-writer-es?43100+57
social-media-agency-pt?2076+56
task-style-no-roleplay-en?48100+52
tax-consultant-en?48100+52
education-tutor-pt?49100+51
education-tutor-es?49100+51
logistics-optimizer-en?4797+50
financial-analyst-en?3786+49
tax-consultant-pt?51100+49
Assistente de leilão judicial de imóveis (pt)pt1665+49
task-style-no-roleplay-pt?52100+48
estrategista-marketing?2876+48
education-tutor-en?53100+47
marketing-strategist-pt?2872+44
Mediación familiar / uniones de hecho (es)es2972+43
tax-consultant-es?5193+42
PMO de hortas hidropônicas em escolas (pt)pt4486+42
tone-domain-mismatch-es?4586+41
logistics-optimizer-pt?5593+38
Satellite ground-station CI reviewer (en)en3876+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
logistics-optimizer-es?63100+37
task-style-no-roleplay-es?65100+35
cooperative-agronomist-es?66100+34
Cybersec bootcamp instructional designer (en)en4175+34
marketing-strategist-es?6493+29
nutripal-weight-loss-en?4876+28
mental-health-wellness-es?5886+28
Cierre y remediación de pasivos mineros (es)es5583+28
contradiction-es?6186+25
Contadora tributária para usinas de etanol (pt)pt4065+25
marketing-strategist-en?4365+22
streaming-script-editor-en?6486+22
contradiction-pt?6179+18
financial-analyst-es?5673+17
contradiction-en?6177+16
threat-framing-es?7993+14
virtual-pharmacist-pt?6579+14
senior-structural-engineer-pt?7790+13
unresolved-reference-en?7686+10
threat-framing-en?8390+7
unresolved-reference-es?7986+7
Pediatric dental clinic intake (en)en3843+5
tone-domain-mismatch-en?9397+4
threat-framing-pt?9093+3
role-inflation-en?9497+3
role-inflation-pt?8886+-2
role-inflation-es?9790+-7
tone-domain-mismatch-pt?9386+-7
unresolved-reference-pt?6958+-11
Command A (Cohere)74 prompts · avg Δ +41.3
promptlangbeforeafterΔ
consultor-juridico?093+93
clinical-nurse-intake-en?090+90
legal-advisor-en?087+87
legal-advisor-pt?086+86
personal-investment-advisor-es?086+86
disguised-repetitions-pt?1390+77
repeticoes-disfarcadas?1390+77
medical-assistant-pt?076+76
ux-research-assistant-en?25100+75
disguised-repetitions-en?2397+74
legal-advisor-es?1486+72
rigid-code-agent-pt?173+72
rigid-code-agent-en?173+72
assistente-medico?072+72
agente-codigo-rigido?173+72
corporate-counsel-pt?071+71
marketing-strategist-pt?2894+66
rigid-code-agent-es?2284+62
Coordinador last-mile en Lima (es)es42100+58
estrategista-marketing?2884+56
medical-assistant-es?2479+55
disguised-repetitions-es?2883+55
tone-domain-mismatch-es?45100+55
Satellite ground-station CI reviewer (en)en3893+55
medical-assistant-en?054+54
social-media-agency-pt?2073+53
education-tutor-es?49100+51
logistics-optimizer-en?4797+50
task-style-no-roleplay-pt?52100+48
Mediación familiar / uniones de hecho (es)es2976+47
task-style-no-roleplay-en?4893+45
Cybersec bootcamp instructional designer (en)en4186+45
Tangier waterfront pier competition (en)en3479+45
tax-consultant-es?5193+42
education-tutor-pt?4990+41
UX lead pre-IPO Avenida Pay (es)es2465+41
Assistente de leilão judicial de imóveis (pt)pt1655+39
tax-consultant-en?4886+38
financial-analyst-en?3775+38
nutripal-weight-loss-en?4886+38
education-tutor-en?5390+37
tax-consultant-pt?5186+35
financial-analyst-pt?3166+35
task-style-no-roleplay-es?65100+35
PMO de hortas hidropônicas em escolas (pt)pt4479+35
Contadora tributária para usinas de etanol (pt)pt4072+32
financial-analyst-es?5687+31
Integridade de risers em águas ultraprofundas (pt)pt6293+31
Pediatric dental clinic intake (en)en3868+30
marketing-strategist-en?4372+29
logistics-optimizer-pt?5583+28
cooperative-agronomist-es?6693+27
contradiction-pt?6186+25
contradiction-es?6186+25
logistics-optimizer-es?6387+24
newsroom-writer-es?4366+23
senior-structural-engineer-pt?77100+23
mental-health-wellness-es?5879+21
marketing-strategist-es?6483+19
threat-framing-en?83100+17
unresolved-reference-en?7693+17
Cierre y remediación de pasivos mineros (es)es5572+17
streaming-script-editor-en?6476+12
contradiction-en?6172+11
virtual-pharmacist-pt?6576+11
threat-framing-es?7986+7
role-inflation-pt?8893+5
role-inflation-en?9497+3
role-inflation-es?97100+3
tone-domain-mismatch-pt?9393+0
unresolved-reference-es?7979+0
threat-framing-pt?9086+-4
tone-domain-mismatch-en?9386+-7
unresolved-reference-pt?6958+-11
Gemini 2.5 Flash74 prompts · avg Δ +39.0
promptlangbeforeafterΔ
consultor-juridico?090+90
agente-codigo-rigido?187+86
rigid-code-agent-pt?186+85
repeticoes-disfarcadas?1397+84
disguised-repetitions-pt?1393+80
legal-advisor-en?079+79
disguised-repetitions-en?23100+77
ux-research-assistant-en?25100+75
rigid-code-agent-en?175+74
medical-assistant-pt?072+72
assistente-medico?069+69
estrategista-marketing?2897+69
UX lead pre-IPO Avenida Pay (es)es2493+69
disguised-repetitions-es?2893+65
rigid-code-agent-es?2286+64
Cybersec bootcamp instructional designer (en)en41100+59
legal-advisor-pt?058+58
legal-advisor-es?1472+58
medical-assistant-en?058+58
corporate-counsel-pt?058+58
personal-investment-advisor-es?057+57
social-media-agency-pt?2076+56
medical-assistant-es?2479+55
Satellite ground-station CI reviewer (en)en3887+49
tone-domain-mismatch-es?4593+48
Coordinador last-mile en Lima (es)es4290+48
education-tutor-en?53100+47
logistics-optimizer-en?4793+46
Assistente de leilão judicial de imóveis (pt)pt1662+46
task-style-no-roleplay-en?4893+45
logistics-optimizer-pt?55100+45
education-tutor-pt?4993+44
clinical-nurse-intake-en?044+44
tax-consultant-pt?5193+42
tax-consultant-en?4890+42
task-style-no-roleplay-pt?5293+41
marketing-strategist-en?4381+38
Integridade de risers em águas ultraprofundas (pt)pt62100+38
Mediación familiar / uniones de hecho (es)es2965+36
tax-consultant-es?5186+35
education-tutor-es?4983+34
cooperative-agronomist-es?66100+34
newsroom-writer-es?4376+33
financial-analyst-pt?3163+32
financial-analyst-en?3768+31
nutripal-weight-loss-en?4879+31
Cierre y remediación de pasivos mineros (es)es5586+31
logistics-optimizer-es?6393+30
PMO de hortas hidropônicas em escolas (pt)pt4472+28
marketing-strategist-pt?2855+27
contradiction-pt?6186+25
contradiction-en?6186+25
task-style-no-roleplay-es?6586+21
virtual-pharmacist-pt?6586+21
mental-health-wellness-es?5879+21
contradiction-es?6179+18
senior-structural-engineer-pt?7793+16
financial-analyst-es?5671+15
streaming-script-editor-en?6479+15
Contadora tributária para usinas de etanol (pt)pt4055+15
threat-framing-en?8397+14
threat-framing-es?7993+14
threat-framing-pt?90100+10
unresolved-reference-en?7686+10
Tangier waterfront pier competition (en)en3444+10
marketing-strategist-es?6472+8
unresolved-reference-es?7986+7
role-inflation-en?94100+6
role-inflation-es?97100+3
tone-domain-mismatch-en?9393+0
Pediatric dental clinic intake (en)en3837+-1
role-inflation-pt?8883+-5
tone-domain-mismatch-pt?9383+-10
unresolved-reference-pt?6958+-11
GPT-5 nano (OpenAI)74 prompts · avg Δ +37.8
promptlangbeforeafterΔ
legal-advisor-en?093+93
rigid-code-agent-pt?193+92
consultor-juridico?087+87
rigid-code-agent-en?181+80
medical-assistant-pt?079+79
personal-investment-advisor-es?079+79
disguised-repetitions-en?23100+77
clinical-nurse-intake-en?073+73
agente-codigo-rigido?171+70
repeticoes-disfarcadas?1383+70
marketing-strategist-pt?2897+69
estrategista-marketing?2897+69
ux-research-assistant-en?2593+68
medical-assistant-en?067+67
financial-analyst-en?3797+60
disguised-repetitions-es?2886+58
assistente-medico?058+58
disguised-repetitions-pt?1370+57
marketing-strategist-en?4397+54
social-media-agency-pt?2074+54
Mediación familiar / uniones de hecho (es)es2983+54
tax-consultant-en?48100+52
education-tutor-pt?49100+51
logistics-optimizer-en?4797+50
task-style-no-roleplay-en?4897+49
UX lead pre-IPO Avenida Pay (es)es2473+49
corporate-counsel-pt?047+47
Assistente de leilão judicial de imóveis (pt)pt1663+47
Satellite ground-station CI reviewer (en)en3884+46
PMO de hortas hidropônicas em escolas (pt)pt4490+46
task-style-no-roleplay-pt?5297+45
legal-advisor-es?1458+44
newsroom-writer-es?4386+43
logistics-optimizer-pt?5597+42
Coordinador last-mile en Lima (es)es4283+41
education-tutor-es?4987+38
Cybersec bootcamp instructional designer (en)en4179+38
logistics-optimizer-es?63100+37
financial-analyst-es?5693+37
legal-advisor-pt?036+36
rigid-code-agent-es?2257+35
task-style-no-roleplay-es?65100+35
mental-health-wellness-es?5893+35
Cierre y remediación de pasivos mineros (es)es5590+35
education-tutor-en?5386+33
streaming-script-editor-en?6497+33
tax-consultant-es?5183+32
Integridade de risers em águas ultraprofundas (pt)pt6293+31
Contadora tributária para usinas de etanol (pt)pt4069+29
cooperative-agronomist-es?6694+28
contradiction-es?6186+25
virtual-pharmacist-pt?6583+18
marketing-strategist-es?6481+17
financial-analyst-pt?3148+17
nutripal-weight-loss-en?4865+17
threat-framing-en?8397+14
unresolved-reference-es?7993+14
medical-assistant-es?2437+13
senior-structural-engineer-pt?7790+13
Tangier waterfront pier competition (en)en3446+12
tax-consultant-pt?5162+11
tone-domain-mismatch-es?4556+11
unresolved-reference-en?7683+7
Pediatric dental clinic intake (en)en3845+7
contradiction-pt?6166+5
contradiction-en?6165+4
role-inflation-pt?8890+2
threat-framing-pt?9090+0
threat-framing-es?7979+0
role-inflation-es?9797+0
tone-domain-mismatch-en?9393+0
role-inflation-en?9487+-7
unresolved-reference-pt?6962+-7
tone-domain-mismatch-pt?9362+-31
GPT-4o mini (OpenAI)74 prompts · avg Δ +35.9
promptlangbeforeafterΔ
legal-advisor-en?085+85
repeticoes-disfarcadas?1397+84
clinical-nurse-intake-en?076+76
rigid-code-agent-en?176+75
disguised-repetitions-en?2394+71
disguised-repetitions-pt?1383+70
agente-codigo-rigido?171+70
legal-advisor-pt?069+69
personal-investment-advisor-es?069+69
rigid-code-agent-pt?168+67
disguised-repetitions-es?2893+65
ux-research-assistant-en?2587+62
medical-assistant-pt?059+59
Mediación familiar / uniones de hecho (es)es2986+57
assistente-medico?056+56
UX lead pre-IPO Avenida Pay (es)es2480+56
medical-assistant-en?055+55
Assistente de leilão judicial de imóveis (pt)pt1669+53
tone-domain-mismatch-es?4597+52
task-style-no-roleplay-en?48100+52
Satellite ground-station CI reviewer (en)en3890+52
marketing-strategist-pt?2878+50
corporate-counsel-pt?050+50
task-style-no-roleplay-pt?52100+48
Coordinador last-mile en Lima (es)es4290+48
social-media-agency-pt?2067+47
logistics-optimizer-en?4793+46
tax-consultant-en?4894+46
nutripal-weight-loss-en?4893+45
education-tutor-es?4990+41
legal-advisor-es?1455+41
consultor-juridico?041+41
Cybersec bootcamp instructional designer (en)en4182+41
education-tutor-en?5393+40
newsroom-writer-es?4383+40
financial-analyst-en?3775+38
Pediatric dental clinic intake (en)en3875+37
logistics-optimizer-pt?5590+35
task-style-no-roleplay-es?65100+35
tax-consultant-pt?5183+32
tax-consultant-es?5183+32
rigid-code-agent-es?2254+32
PMO de hortas hidropônicas em escolas (pt)pt4476+32
education-tutor-pt?4979+30
logistics-optimizer-es?6393+30
Integridade de risers em águas ultraprofundas (pt)pt6290+28
cooperative-agronomist-es?6693+27
Contadora tributária para usinas de etanol (pt)pt4065+25
medical-assistant-es?2448+24
Tangier waterfront pier competition (en)en3458+24
contradiction-es?6183+22
virtual-pharmacist-pt?6586+21
mental-health-wellness-es?5879+21
financial-analyst-pt?3151+20
Cierre y remediación de pasivos mineros (es)es5572+17
marketing-strategist-en?4358+15
role-inflation-pt?88100+12
streaming-script-editor-en?6476+12
unresolved-reference-en?7686+10
estrategista-marketing?2837+9
contradiction-pt?6169+8
threat-framing-en?8390+7
threat-framing-es?7986+7
role-inflation-en?94100+6
senior-structural-engineer-pt?7783+6
financial-analyst-es?5661+5
contradiction-en?6166+5
marketing-strategist-es?6467+3
threat-framing-pt?9093+3
role-inflation-es?97100+3
unresolved-reference-pt?6972+3
tone-domain-mismatch-pt?9393+0
tone-domain-mismatch-en?9393+0
unresolved-reference-es?7979+0

Represent an AI model or provider and want to see how it performs here?