RankingEffort benchmarks

Claude Haiku 5.5

Compare models

121 measured results · 5 effort settings · 38 benchmark/harness combinations

Model evidence and sources →

Change model

Model

Current configurationSources

Compare with

Choose a second model and its effort to see both profiles and benchmark differences below.

Compare withModel and configuration

Capability profile

One configuration, seven capabilities

Comparison score · 0–100

AClaude Haiku 5.5 · Max4 of 7 capabilities supported

Gaps are unknown, not zero. Hollow points are preliminary.

Agentic
A43.7Supported
Hard reasoning
A42.9Supported
Coding
A64.4Supported
Human pref
Unknown
Knowledge
A14.6Supported
Multimodal
Unknown
Long context
A66.8Preliminary

Each profile belongs to the named configuration. Capability scores use a shared panel for all axes, distinct from the overall ranking panel; they are not accuracy percentages or direct head-to-head wins.

Capability score evidence

Claude Haiku 5.5 · Max: 48.2 overall comparison score. Meets the overall evidence requirements.

5 areas measured. The available task mix affects this estimate; missing areas remain unknown.

Which sources carry the weight?

Cognition · 48.6% direct comparison weight · independent of the model provider
Artificial Analysis · 29.9% direct comparison weight · independent of the model provider
LiveBench · 21.5% direct comparison weight · independent of the model provider

Which benchmark families carry the weight?

FrontierCode · 48.6% direct comparison weight
Humanity’s Last Exam · 3.7% direct comparison weight
Critpt · 3.7% direct comparison weight
omniscience · 3.6% direct comparison weight
Aa Lcr · 3.4% direct comparison weight
Other families (14) · 37.0% direct comparison weight

These shares describe direct comparison weight, not fractions of the final score or confidence. Source balancing does not give every family equal weight or remove correlations between overlapping sources.

How sensitive is this score?

The audit removes one source at a time and refits against the same overall reference panel. These published results use the matching dataset audited 2026-10-08.

Both ranks use the same surviving configurations for each removal. The cohort starts with 444 configurations supported in the original audit, including separate effort settings. These are diagnostic positions, not the default leaderboard ranks. Eligibility is not rechecked after removal.

What changes when a source is removed?
Removed source	Rank before → after	Rank movement	Cohort remaining / lost support	Score after (change)
Artificial Analysis	62 → 65	Down 3	204 remaining / 240 lost	46.9 (-1.3 points)
ARC Prize	65 → 70	Down 5	444 remaining / 0 lost	49.6 (+1.4 points)
Datacurve	65 → 67	Down 2	444 remaining / 0 lost	45.4 (-2.8 points)
LiveCodeBench	65 → 65	No movement	444 remaining / 0 lost	48.2 (-0.0 points)
Scale	65 → 65	No movement	444 remaining / 0 lost	49.7 (+1.5 points)
Terminal-Bench	65 → 65	No movement	444 remaining / 0 lost	48.3 (+0.2 points)
Cognition	65 → 90	Down 25	444 remaining / 0 lost	36.6 (-11.6 points)
LiveBench	65 → 56	Up 9	443 remaining / 1 lost	58.5 (+10.3 points)
SWE-rebench	65 → 66	Down 1	444 remaining / 0 lost	47.7 (-0.5 points)
Epoch AI	64 → 66	Down 2	426 remaining / 18 lost	50.1 (+1.9 points)
Vals AI	65 → 67	Down 2	443 remaining / 1 lost	46.4 (-1.7 points)
Provider reports · Anthropic	65 → 65	No movement	444 remaining / 0 lost	48.2 (-0.0 points)
Provider reports · Moonshot AI	65 → 65	No movement	444 remaining / 0 lost	47.9 (-0.3 points)
Provider reports · Meta	65 → 66	Down 1	444 remaining / 0 lost	48.0 (-0.1 points)
Provider reports · Mistral	65 → 65	No movement	444 remaining / 0 lost	48.2 (+0.0 points)
Provider reports · OpenAI	65 → 66	Down 1	444 remaining / 0 lost	48.2 (+0.1 points)
All provider reports	65 → 67	Down 2	444 remaining / 0 lost	47.8 (-0.4 points)

Lost support means a configuration no longer has a comparison path to the complete reference panel. Its rank is unavailable. A surviving path alone does not establish leaderboard eligibility.

Overlapping sources may remain correlated even with source balancing. Leave-one-source-out movement is a sensitivity diagnostic, not a spread of source-specific ranks or a confidence measure. Scoring methodology · Dated configuration audit

Every benchmark family

Removed evidence	Recomputed score	Change
frontiercode	36.6	-11.6
deepswe	45.4	-2.8
multi-swe-bench	45.9	-2.3
livebench-instructions	50.2	+2.0
livebench-language	49.5	+1.3
mmmu-pro	49.4	+1.3
omniscience	49.3	+1.1
arc-agi	49.1	+0.9
terminal-bench-science	47.3	-0.8
livebench-coding	47.4	-0.8
briefcase	47.4	-0.8
automationbench	48.9	+0.7
hle	48.8	+0.7
swe-rebench	47.7	-0.5
aime	47.7	-0.4
livebench-reasoning	47.7	-0.4
livebench-data	48.6	+0.4
arc-agi-3	47.8	-0.4
mmlu-pro	48.5	+0.4
enigma-eval	48.5	+0.4
ifbench	47.9	-0.3
gdpval	47.9	-0.3
swe-atlas-qna	47.9	-0.3
epoch-game-puzzles	48.5	+0.3
swe-atlas-test-writing	47.9	-0.3
simpleqa	48.4	+0.2
critpt	48.4	+0.2
gpqa	48.4	+0.2
livebench-math	48.0	-0.2
tau-bench	48.0	-0.2
vibe-code	48.0	-0.2
gdp-pdf	48.3	+0.1
frontiermath	48.0	-0.1
harvey	48.0	-0.1
apex-agents	48.3	+0.1
vals-finance-agent	48.1	-0.1
vals-excel-modeling	48.1	-0.1
aa-lcr	48.2	+0.1
itbench	48.2	+0.1
vals-legal-research	48.1	-0.1
terminal-bench	48.2	+0.1
livecodebench	48.1	-0.1
vals-tax-agent	48.1	-0.1
enterprise-ops	48.2	+0.1
swe-atlas-refactoring	48.1	-0.1
analyst-agent	48.2	+0.0
math500	48.2	-0.0
vals-code-migration	48.2	-0.0
swe-bench-pro	48.2	+0.0
gmmlu	48.2	-0.0

What evidence is missing from the fit?

Across all collected settings for this model: 177 observations, 111 contributing, 0 matched without graph weight and 66 excluded. Counts do not establish rank eligibility.

5 · Benchmark family or source protocol has not been reviewed
5 · Original SciCode grading defects are under review; retained as supporting evidence, excluded from aggregate and capability scores. Corrected SciCode-Verified requires a separate protocol review. Sources: https://arxiv.org/abs/2608.04975 and https://artificialanalysis.ai/evaluations/scicode
53 · Provider publication; exact benchmark-specific configuration and independent evaluation provenance require separate review before matched-board admission.
3 · Supporting evidence outside the reviewed capability core

Inspect every observation and exclusion →

Benchmark results

These are published benchmark scores. Supported effort estimates also appear on the Capability leaderboard. Results from different harnesses occupy different rows. Missing cells stay unknown; no scores are copied between effort levels.

Benchmark / source	Low	Medium	High	XHigh	Max
AA-Briefcase EloArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Briefcase v1.1 is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-08.	1,111.58 Elo Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	1,371.7 Elo Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	1,442.55 Elo Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	1,532.67 Elo Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	1,577.72 Elo Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
AA-LCR v1.1Artificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	72% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	77.33% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	77.33% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	78.33% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	82.67% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
AutomationBench-AAArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of task objectives completed with no guardrail violations · Higher is better · Benchmark developed by Zapier · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	22.91% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	28.59% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	33.73% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	35.95% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	35.41% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
CritPtArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark developed by Argonne and UIUC, with contributions from 60+ researchers globally. CritPt is under review by its authors following external feedback · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	9.14% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	12.86% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	18.57% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	22.57% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	18.86% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
GDP.pdf: All-pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of attempts where every atomic criterion passed · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	11.2% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	15.2% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	17.2% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	18.2% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	20.8% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
GDPval-AA v2.1 LeaderboardArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Elo rating for performance on real-world work tasks · Anchored to DeepSeek V4.1 Flash (max) at 1600 · Higher is better · Evaluation results measured independently by Artificial Analysis Unit: elo. Source collected 2026-10-08.	1,124.94 Elo Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	1,276.84 Elo Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	1,420.03 Elo Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	1,510.76 Elo Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	1,620.05 Elo Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
Harvey LAB-AA: Criterion Pass RateArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Share of atomic pass/fail rubric criteria the deliverables satisfy, graded by a single LLM judge · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	87.86% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	89.23% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	89.52% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	89.9% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	89.87% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
Humanity's Last ExamArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	27.02% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	33.78% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	37.26% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	42.68% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	44.39% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
AA-Omniscience IndexArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct. · Evaluation results measured independently by Artificial Analysis Unit: index. Source collected 2026-10-08.	3.25 Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	4.47 Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	5.75 Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	6.05 Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	10.67 Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
AA-Omniscience AccuracyArtificial Analysis · 6000 questions; 1 attempt; correct-answer share; GPT-5.6 Luna medium grader Benchmark details AA-Omniscience Accuracy (higher is better) measures the proportion of correctly answered questions out of all questions, regardless of whether the model chooses to answer · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	33.25% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	34.03% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	34.78% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	34.95% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	36.38% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
SciCodeArtificial Analysis · Artificial Analysis published evaluation; see source for benchmark-specific settings Benchmark details Benchmark curated by scientists across 16 disciplines. SciCode is under review following an independent audit of the dataset · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	49.19% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	48.96% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	48.73% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	51.74% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	54.98% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
Terminal-Bench-Science 0.1Artificial Analysis · mini-swe-agent; Terminal-Bench-Science 0.1.0; 70 tasks; pass@1 averaged over 3 repeats Benchmark details Share of tasks passed (pass@1), averaged over three repeats · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	1.9% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	7.14% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	10% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	18.1% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	20% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
Terminal-Bench v4.0Artificial Analysis · mini-SWE-agent v2.4.6; 66 tasks; pass@1 averaged over 3 repeats; 500 steps; task timeout up to 8 hours Benchmark details Benchmark developed by the Laude Institute, Stanford, and open-source contributors · Independently benchmarked by Artificial Analysis · Evaluation results measured independently by Artificial Analysis Unit: percent. Source collected 2026-10-08.	12.63% Reported settings Claude Haiku 5.5 (Low); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-low	15.15% Reported settings Claude Haiku 5.5 (Medium); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-medium	21.72% Reported settings Claude Haiku 5.5 (High); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-high	29.29% Reported settings Claude Haiku 5.5 (Xhigh); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5-xhigh	32.83% Reported settings Claude Haiku 5.5 (Max); Model source: https://artificialanalysis.ai/models/claude-haiku-5-5
FrontierCode 1.1 · extended-claude-codeCognition · claude-code Benchmark details FrontierCode 1.1; claude-code. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-08.	48.19% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	55.04% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	55.88% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	58.12% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort xhigh; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	58.43% Reported settings FrontierCode 1.1; extended; 150 tasks; claude-code; source effort max; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified
FrontierCode 1.1 · main-claude-codeCognition · claude-code Benchmark details FrontierCode 1.1; claude-code. Open each result for its exact effort, task subset, metric and protocol details. Unit: percent. Source collected 2026-10-08.	34.76% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort low; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	41.63% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort medium; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	41.89% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort high; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	45.83% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort xhigh; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified	46.36% Reported settings FrontierCode 1.1; main; 100 tasks; claude-code; source effort max; weighted rubric score, not pass rate; unfair internet runs zeroed by evaluator; flag rate unspecified
LiveBench · AMPS Hard · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	98% Reported settings LiveBench official task evaluation; livebench-2026-06-25-AMPS_Hard; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	97% Reported settings LiveBench official task evaluation; livebench-2026-06-25-AMPS_Hard; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · code completion · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	73.91% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_completion; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	78.26% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_completion; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · code generation · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	78.87% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_generation; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	77.47% Reported settings LiveBench official task evaluation; livebench-2026-06-25-code_generation; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · connections · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	93.83% Reported settings LiveBench official task evaluation; livebench-2026-06-25-connections; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	96% Reported settings LiveBench official task evaluation; livebench-2026-06-25-connections; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · consecutive events · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	71.05% Reported settings LiveBench official task evaluation; livebench-2026-06-25-consecutive_events; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	10.64% Reported settings LiveBench official task evaluation; livebench-2026-06-25-consecutive_events; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · integrals with game · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	89% Reported settings LiveBench official task evaluation; livebench-2026-06-25-integrals_with_game; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	67% Reported settings LiveBench official task evaluation; livebench-2026-06-25-integrals_with_game; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · javascript · 2026-06-25LiveBench · LiveBench Mini-SWE-Agent, 250-step limit Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	59.09% Reported settings LiveBench Mini-SWE-Agent, 250-step limit; livebench-2026-06-25-javascript; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	68.18% Reported settings LiveBench Mini-SWE-Agent, 250-step limit; livebench-2026-06-25-javascript; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · logic with navigation · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	56% Reported settings LiveBench official task evaluation; livebench-2026-06-25-logic_with_navigation; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	66% Reported settings LiveBench official task evaluation; livebench-2026-06-25-logic_with_navigation; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · math comp · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	97.06% Reported settings LiveBench official task evaluation; livebench-2026-06-25-math_comp; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	98.04% Reported settings LiveBench official task evaluation; livebench-2026-06-25-math_comp; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5
LiveBench · olympiad · 2026-06-25LiveBench · LiveBench official task evaluation Benchmark details LiveBench 2026-06-25 task snapshot. Task scores stay separate and related tasks share their reviewed family budget. Explicit checkpoint and effort labels only. Unit: percent. Source collected 2026-10-08.	Not reported	Not reported	Not reported	86.95% Reported settings LiveBench official task evaluation; livebench-2026-06-25-olympiad; claude-haiku-5-5-xhigh-effort; checkpoint=claude-haiku-5-5	83.7% Reported settings LiveBench official task evaluation; livebench-2026-06-25-olympiad; claude-haiku-5-5-max-effort; checkpoint=claude-haiku-5-5

What these comparisons mean

Each source reports its own effort labels. Matching labels do not establish equal compute budgets or identical fallback behavior. Open Reported settings to inspect the exact model name, agent and task coverage. Multiple results at one level remain visible rather than selecting the best. Source-reported None or Non-reasoning labels are not treated as evidence that an API supports those settings.

These measurements use their own dated source collection. The leaderboard fits identified configurations jointly; it contains no pooled model ratings. Explicit thinking budgets are separate configurations. Read the methodology.

Download all effort benchmark results · 8,798 results across 256 models · Collected 2026-10-08

Claude Haiku 5.5

Compare with

One configuration, seven capabilities

Agentic

Hard reasoning

Coding

Human pref

Knowledge

Multimodal

Long context

Capability score evidence

Benchmark results