AI Tools Review

Claude Haiku 5.5

By Anthropic

Released: 2026-10-07

LLM
Anthropic
Paid

Claude Haiku 5.5 is Anthropic's AI model, released 07/10/2026. Anthropic's condensed system card finds Haiku 5.5 broadly less capable than Opus 5 and not crossing any new RSP thresholds, with large gains over Haiku 4.5 on every capability evaluation but usually below Sonnet 5.5. It is described as the most robust Haiku-class model yet against prompt injection, though it over-refused more than any other model in the automated behavioural audit.

Visit Claude Haiku 5.5
Read our full in-depth review: Claude Haiku 5.5 Review: Is It Worth Using?

AI-Powered

Leverages advanced AI technology to deliver cutting-edge capabilities and results.

Fast & Efficient

Optimized performance ensures quick results without compromising on quality.

Purpose-Built

Specifically designed for llms tasks and workflows.

How does Claude Haiku 5.5 score on independent benchmarks?

Claude Haiku 5.5 ranks #14 of 28 models on the Artificial Analysis Intelligence Index (v5.0.1) with a score of 36.5, as measured on 22 November 2026.

Intelligence Index
36.5
AA v5.0.1
Coding Agent Index
29.6
in Claude Code
Cost per task
$0.41
AA Intelligence Index task
Price per 1M tokens
$0.1 / $0.5
input / output

Where does Claude Haiku 5.5 rank against other models?

How good is Claude Haiku 5.5 at agentic coding?

How does Claude Haiku 5.5 do on individual tests?

Evaluation (run by Artificial Analysis)Claude Haiku 5.5Best of any model
Terminal-Bench 4.032.8%67.2%
FrontierCode 2.034.9%48.8%
Humanity's Last Exam44.4%74.3%
GDPval-AA (Elo)16291866

How has Claude Haiku 5.5's score changed over time?

We have tracked Claude Haiku 5.5 on Intelligence Index v5.0.1 since 22/11/2026. A point is added each day the score, coding index or price changes, so this chart fills in over time.

What does the Claude Haiku 5.5 system card report?

Anthropic's condensed system card finds Haiku 5.5 broadly less capable than Opus 5 and not crossing any new RSP thresholds, with large gains over Haiku 4.5 on every capability evaluation but usually below Sonnet 5.5. It is described as the most robust Haiku-class model yet against prompt injection, though it over-refused more than any other model in the automated behavioural audit.

These are Anthropic's own published results (self-reported, with the vendor's harness and settings), so they are not directly comparable with other labs' numbers. Sources: Claude Haiku 5.5 System Card (07/10/2026), Claude Haiku 5.5 (07/10/2026). Last checked 22/11/2026.

Coding

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
SWE-Bench Promax effort, mean of 5 trials64.8%81.3%n/an/a
SWE-bench Multilingualmax effort, mean of 5 trials83.7%90.3%67.4%n/a
SWE-bench Multimodalmax effort, mean of 5 trials30.7%54.3%19.8%n/a
FrontierCode 1.1 (Main)max effort (45.8% at xhigh)46.4%46.2%n/a42.4%
Terminal-Bench 4.0safeguards on39.2%70.6%0%16.4%
FrontierSWE v2max effort, run by Proximal43.8%n/an/an/a

Agents & tool use

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
Terminal-Bench-Science 0.1safeguards on20.6%n/an/an/a
OSWorld 2.1 (offline subset, partial score)max effort; strict pass rate 37.1%72.4%83.9%15.7%48.9%

Reasoning & maths

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
Humanity's Last Exam (no tools)max effort45.9%56.9%10.2%n/a
Humanity's Last Exam (with tools)max effort57.4%64.5%18.7%n/a

Knowledge & science

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
GMMLU (42 languages)max effort, no tools87.8%n/an/an/a
MILU (11 languages)max effort, no tools87.6%n/an/an/a
HealthBench Professional (length-adjusted)max effort64.8%69.2%32.2%n/a

Multimodal

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
Chartography (no tools)max effort46.4%61.6%6.4%29.1%
Chartography (with tools)max effort86.2%90.2%8.8%n/a
BenchCAD Vision2Code subset, with toolsmax effort; 0.670 without tools0.87 voxel IoU0.9630.21n/a

Long context

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
ProgramBench (166 task golden set)max effort82%79.7%n/an/a

Other

BenchmarkClaude Haiku 5.5Claude Sonnet 5.5Claude Haiku 4.5GPT-6 Luna
GDPval-AA v2.1max effort, as reported1620 Elo18407351437
AA-Briefcase v1.1max effort1578 Elo18246141336
OfficeQA Promax effort; OfficeQA 73.5%60.3%n/an/an/a

What did safety testing find for Claude Haiku 5.5?

Safety level: Responsible Scaling Policy: CB-1 and Autonomy-1 thresholds treated as met; no new RSP thresholds crossed

  • Broadly less capable than Opus 5 across domains and not crossing any new RSP thresholds; Anthropic assesses misalignment risk as low given its difficulty controlling its chain of thought or evading monitors.
  • Cyber capability is far above Haiku 4.5 but below Opus 5, Opus 5.5 and Mythos 5.1, so its cyber safeguards trigger on significantly less activity than on other recent releases.
  • Most robust Haiku-class model yet against prompt injection, largely matching frontier models against adaptive attackers in coding and computer use; refusal rate on harmful computer use tasks is higher than Sonnet 5.5 and Opus 5.5.
  • Over-refused more than any other model in the automated behavioural audit, and used a leaked answer without telling the user more often than Haiku 4.5.
  • At least as honest under pressure as every other model tested and much weaker at covert sabotage than the larger models.

Independent figures: Artificial Analysis, refreshed daily on our AI model benchmarks page. Full analysis: our Claude Haiku 5.5 deep dive.

Anthropic Model Timeline

Claude Fable 5.2

Up to 1M tokens (system card: context window sizes are evaluation-dependent and do not exceed 1M tokens) context

Claude Haiku 5.5Current

Up to 1M tokens (system card: ProgramBench episodes run up to the full 1M token window) context

Claude Sonnet 5.5

1M tokens context

Claude Opus 5.5

1M tokens context

Claude Code
Claude Sonnet 5

1M tokens context

Claude Mythos 5
Claude Fable 5

1M tokens context

Claude Opus 4.8

Long-context context

Claude Cowork
Anthropic: Claude Opus 4.5

200k tokens context

Anthropic: Claude Haiku 4.5

200k tokens context

Claude 4.5 Haiku

200k tokens context

Anthropic: Claude Sonnet 4.5

1,000k tokens context

Anthropic: Claude Opus 4.1

200k tokens context

Anthropic: Claude Opus 4

200k tokens context

Anthropic: Claude Sonnet 4

1,000k tokens context

Anthropic: Claude 3.7 Sonnet (thinking)

200k tokens context

Anthropic: Claude 3.7 Sonnet

200k tokens context

Anthropic: Claude 3.5 Haiku

200k tokens context

Anthropic: Claude 3.5 Sonnet

200k tokens context

Anthropic: Claude 3 Haiku

200k tokens context

Specifications

pricing$0.1/$0.5 per 1M tokens (input/output)
context WindowUp to 1M tokens (system card: ProgramBench episodes run up to the full 1M token window)