Request a pilot
‹ PublicAI Index

The LLM benchmark aggregator.

Claude Haiku 4.5

Anthropic

Strongest in Tool use (#6 of 81), weakest in Secure code (#194 of 274). Above par in 11 of 27 scopes. Among the models it meets almost everywhere, it finishes behind Claude Fable 5 and Claude Fable 5.1 and ahead of DeepSeek R1 and Grok 3.

§ 1 · Profile

What it is good at

Bars run from 50 — the average of the models each source lists — so right of the line is above par. The middle column is the gap to whoever leads that scope.

Safety55.4#44/337
Jailbreak resistance61.7#26/272
Harm refusal57.6#27/300
Toxicity avoidance57.4#42/272
Fairness57#50/300
Factual grounding50.4#56/101
Safe-prompt compliance47.9#61/82
Secure code45.1#194/274
Agents55#65/268
Tool use72.4#6/81
Knowledge work47.3#94/178
Reasoning49.1#94/178
Mathematics54.4#59/140
Science47.3#84/122
Reasoning43.6#89/140
Knowledge47.4#101/138
Academic knowledge46.9#93/123
Human preference58.4#109/342
Human preference58.4#109/342
Professional46#121/168
Medical47.2#95/140
Finance46.3#106/151
Legal44.2#107/151
Core abilities46.3#135/204
General intelligence44.7#137/204
Coding41.9#143/165
Agentic coding40.9#137/157

§ 2 · Head to head

What it beats, and what beats it

The same models turn up scope after scope. Counted once: where both were placed, who finished higher. Bars run right for Claude Haiku 4.5, left for the other.

Mistral Smallwon 20 · lost 1
Gemini 1.5 Prowon 18 · lost 1
Gemini 1.5 Flashwon 16 · lost 1
GPT-4o Miniwon 22 · lost 2
Command R Pluswon 16 · lost 2
GPT-4.1 Nanowon 23 · lost 3
Mistral Medium 3.5won 15 · lost 2
Llama 4 Scout Instructwon 22 · lost 3
Mistral Largewon 16 · lost 3
GPT OSS 20Bwon 18 · lost 4
Command Awon 19 · lost 5
Gemini 2.0 Flashwon 17 · lost 6
Mistral Large 3won 14 · lost 6
Claude 3.5 Haikuwon 13 · lost 7
Gemini 2.5 Flash Litewon 13 · lost 7
GPT-4owon 14 · lost 8
GPT-5 Nanowon 15 · lost 9
GPT-4.1 Miniwon 16 · lost 10
Ling 3.0 Flashwon 11 · lost 7
O3 Miniwon 14 · lost 9
Grok 4.3won 14 · lost 10
Grok 3won 11 · lost 10
DeepSeek R1won 13 · lost 12
DeepSeek V3won 13 · lost 13
GPT OSS 120Bwon 12 · lost 14
O4 Miniwon 12 · lost 14
Kimi K2won 12 · lost 14
Gemma 4 31Bwon 10 · lost 12
GPT-5.4 Nanowon 9 · lost 11
Qwen3.6 Pluswon 8 · lost 10
MiniMax M3won 11 · lost 14
Gemini 2.5 Flashwon 11 · lost 15
GLM-4.5won 8 · lost 11
Gemini 3.1 Flash Litewon 8 · lost 11
GPT-5.4 Miniwon 8 · lost 11
Kimi K2.6won 10 · lost 15
Grok 4.1 Fastwon 9 · lost 14
Mimo V2.5 Prowon 9 · lost 14
Claude 3.7 Sonnetwon 8 · lost 14
Grok 4won 9 · lost 16
Qwen3 235B A22B Instructwon 9 · lost 16
GPT-4.1won 9 · lost 17
Grok 4 Fastwon 6 · lost 12
Gemini 3.5 Flash Litewon 6 · lost 12
Kimi K2.5won 8 · lost 17
GPT-5.6 Lunawon 8 · lost 17
GLM-5.1won 7 · lost 15
Claude Sonnet 4won 8 · lost 18
DeepSeek V3.2won 8 · lost 18
Claude 3.5 Sonnetwon 6 · lost 14
GLM-5.3 Flashwon 7 · lost 17
Qwen3.7 Maxwon 7 · lost 17
Gemini 3.5 Flashwon 7 · lost 17
O1won 6 · lost 15
MiniMax M2.7won 6 · lost 15
GPT-6 Lunawon 6 · lost 15
GLM-5.2won 7 · lost 18
Mimo V2.6 Prowon 5 · lost 14
GPT-5 Miniwon 7 · lost 20
Kimi K2 Thinkingwon 5 · lost 15
Gemini 3.1 Prowon 5 · lost 15
DeepSeek V4 Flashwon 6 · lost 18
Grok 4.5won 6 · lost 18
Qwen3.8 Maxwon 6 · lost 18
Gemini 3.6 Flashwon 6 · lost 18
Gemini 3.8 Flashwon 6 · lost 18
GPT-5.6 Solwon 6 · lost 19
Gemini 2.5 Prowon 6 · lost 20
GPT-6 Solwon 5 · lost 17
Claude Sonnet 4.6won 5 · lost 18
Qwen3.8 27Bwon 4 · lost 15
Claude Opus 4won 5 · lost 19
GLM-4.6won 5 · lost 20
GPT-5.2won 5 · lost 20
Inklingwon 5 · lost 20
DeepSeek V4 Prowon 5 · lost 20
GPT-5.4won 5 · lost 20
Claude Opus 4.6won 4 · lost 17
Grok 4.6won 4 · lost 20
GPT-5.1won 4 · lost 20
O3won 4 · lost 21
GPT-5.6 Terrawon 4 · lost 21
GPT-5.5won 4 · lost 21
Claude Sonnet 4.5won 4 · lost 23
Grok 4.7won 3 · lost 18
Gemini 3 Prowon 3 · lost 20
Claude Opus 5won 3 · lost 21
Claude Opus 4.7won 3 · lost 22
DeepSeek V4.1 Flashwon 2 · lost 15
GPT-5won 3 · lost 23
GLM-5won 2 · lost 16
Claude Opus 4.5won 2 · lost 18
Claude Opus 5.5won 2 · lost 19
Muse Spark 1.1won 2 · lost 22
GPT-6 Astrawon 1 · lost 16
Muse Spark 1.2won 1 · lost 16
GLM-5.3won 1 · lost 17
Kimi K3won 1 · lost 17
Claude Opus 4.8won 1 · lost 23
Claude Sonnet 5won 1 · lost 24
Gemini 3.7 Flashwon 0 · lost 18
Claude Fable 5.1won 0 · lost 18
Claude Fable 5won 0 · lost 18

§ 3 · Sources

Where the numbers come from

10 publications, 35 figures. Every one links to the page it was read from.

LMArena Text ↗1 measureread 2026-09-26
LMArena Text 1414
Artificial Analysis ↗1 measureread 2026-09-26
Artificial Analysis Intelligence Index 17
GDPval-AA ↗1 measureread 2026-09-26
GDPval-AA 719
AA-Briefcase ↗1 measureread 2026-09-26
AA-Briefcase 616
ARC-AGI-2 ↗1 measureread 2026-09-26
ARC-AGI-2 4%
BFCL ↗1 measureread 2026-09-26
BFCL v4 68.7%
Vals.ai ↗17 measuresread 2026-09-26
Vals · Legal Research Bench 10.58%Vals · CaseLaw 56.48%Vals · LegalBench 81.24%Vals · Harvey Legal Agent Benchmark 0.83%Vals · Finance Agent 31.01%Vals · CorpFin 60.61%Vals · TaxEval 67.54%Vals · MortgageTax 62.16%Vals · MedQA 79.57%Vals · MedCode 32.68%Vals · MedScribe 85.23%Vals · SWE-bench Verified 66.6% (Mini-SWE-agent)Vals · Vibe Code Bench 11.39% (OpenHands)Vals · Code Migration 7.55%Vals · GPQA Diamond 72.22%Vals · MMLU Pro 78.72%Vals · AIME 82.71%
HELM Safety ↗5 measuresread 2026-09-26
HELM Safety · HarmBench 95.9%HELM Safety · SimpleSafetyTests 98.8%HELM Safety · Anthropic Red Team 96.9%HELM Safety · BBQ 92.8%HELM Safety · XSTest 93.2%
Enkrypt AI Safety Leaderboard ↗6 measuresread 2026-09-26
Enkrypt · Jailbreak risk 2.9%Enkrypt · Harmful content risk 1.1%Enkrypt · CBRN risk 10.5%Enkrypt · Toxicity risk 1.1%Enkrypt · Bias risk 64.9%Enkrypt · Insecure code risk 43.6%
Vectara · Factual consistency 90.2%

Badge

PublicAI Index badge for Claude Haiku 4.5[![PublicAI Index](https://publicai.io/model-index/badge?model=claude-haiku-4-5)](https://publicai.io/model-index/m/claude-haiku-4-5)