Claude 3 Haiku

Claude 3 Haiku is a non-reasoning model from Anthropic. 13 benchmarks count toward its score, in 7 categories.

availableShows if the model has enough results for an index.
IndexOverall score. 50 is the middle.21.1 ±5.0
CoverageShare of the index weight with results.95%
SpeedOutput tokens per second.73/s
Input / 1MUS dollars per 1M input tokens.$0.25
Output / 1MUS dollars per 1M output tokens.$1.25
ContextMaximum tokens in one request.200K
EloLMArena rating and rank.1195 (#307)

50 is the middle of the board. The range shows the doubt in the index.

117,701 votes. Elo shows what people prefer. It does not change the score.

CapabilitiesScore per category. 50 is the middle.

50 is the middle
AgenticMulti-step tasks with tools.
14.0
CodingCode writing and repair.
30.1
ReasoningLogic problems and puzzles.
32.9
MultimodalTasks with images and text.
5.0
KnowledgeFacts and expert knowledge.
22.5
MultilingualTasks in many languages.
N/A
InstructionTasks with strict rules in the prompt.
25.8
MathMath problems.
9.3

Results

13 counted
BenchmarkThe test name.CategoryThe capability that the test measures.ResultThe score from the publisher.IndexThis result on the index scale.RunThe settings of the run.DateDate of the result.Published byThe source of the result.
SWE-bench VerifiedCoding40.6%30.1Tools1 Sept 2026SWE-bench team
Artificial Analysis GPQA DiamondKnowledge37.4%6.9Artificial Analysis
GPQA diamondKnowledge36.3%11.8Epoch AI
Artificial Analysis IFBenchInstruction36.1%25.8Artificial Analysis
Artificial Analysis MMMU-ProMultimodal30.8%5.0Artificial Analysis
Artificial Analysis Long Context ReasoningReasoning27.7%27.4Artificial Analysis
τ²-Bench Tool-Agent-User EvaluationAgentic21.1%14.0Victor Barres et al.
Artificial Analysis Omniscience AccuracyKnowledge17.6%35.6Artificial Analysis
MATH level 5Math14.9%5.9Epoch AI
Artificial Analysis Intelligence IndexKnowledge5.6%29.4Artificial Analysis
Artificial Analysis Humanity's Last ExamKnowledge4.1%29.0Artificial Analysis
OTIS Mock AIME 2024-2025Math1.8%12.7Epoch AI
Critical Physics TasksReasoning0.0%38.4Artificial Analysis

13 benchmarks count, from 13 of 13 results. A grey row does not count. Too few models took that benchmark.

Sources

BenchLM benchmark aggregationUsed with attribution; per-benchmark results credited to their original authorsOpenRouter, collected directlyNo licence statedSWE-bench team, collected directlyNo licence stated. The repository publishes submission records for reproducibility and transparency and asks that SWE-bench be citedEpoch AI, collected directlyCC BY — free to use and redistribute with attribution

Same level, lower price

Qwen3.8-Flash-Next66.4 · Freedots3-note Preview65.5 · FreeApodex 1.1 Mini59.1 · Free

More from Anthropic

Claude Opus 5.582.8Claude Fable 5.181.4Claude Opus 577.9Claude Fable 577.5Claude Opus 4.871.1