Claude 3.5 Haiku
Claude 3.5 Haiku is behind the leaders in multimodal tasks and instruction following. Too few results yet to rate reasoning, coding, agentic tasks, safety, long context, math, multilingual tasks, or factuality.
Capability profile
Bars show the model's median result as a share of the leading model's, per capability. Select a row to see its results.
Too few results yet to rate Reasoning, Coding, Agentic, Safety, Long Context, Math, Multilingual or Factuality.
Price
$0.80input$4.00outputper million tokens
From Anthropic's own price page · All prices
Evidence
46results on37benchmarks
- 14 independently verified
- 18 aggregator
- 8 vendor-reported
- 6 cross-referenced
From 12 sources · latest Oct 8, 2026 · How verification works
API features
Tool calling
As listed by OpenRouter
Claude 3.5 Haiku benchmark results
46 results on 37 benchmarks, grouped by capability. Each bar is the result as a share of the capability leader's; every result links to its source.
41.1% behind the leader1 of 6 ranked benchmarks measured
- MMMU-Pro45.61Oct 8, 2026aa_mmmu_pro
43.1% behind the leader1 of 3 ranked benchmarks measured
- IFBench42.79Oct 8, 2026aa_ifbench
Show 2 more instruction following resultsHide 2 instruction following results
- LiveBench · Instruction Following66.20Aug 23, 2026livebench_instruction_following@2025-04-07
- LiveBench · Instruction Following65.77Jun 17, 2026livebench_instruction_following@2025-04-07
0 of 6 ranked benchmarks measured
- 0.00Oct 8, 2026
- GPQA Diamond40.81Oct 8, 2026gpqa
- Humanity's Last Exam3.63Oct 8, 2026aa_hle
Show 2 more reasoning resultsHide 2 reasoning results
- GPQA Diamond41.60Oct 7, 2026GPQA
- 41.60Aug 25, 2026
0 of 10 ranked benchmarks measured
- LiveBench · Coding50.78Aug 23, 2026livebench_coding@2025-04-07
- LiveBench · Coding50.78Jun 17, 2026livebench_coding@2025-04-07
- Terminal-Bench 2.110.11Oct 8, 2026terminalbenchV21
Show 3 more coding resultsHide 3 coding results
- SciCode27.43Sep 4, 2026aa_scicode
- 40.60Oct 7, 2026
- Terminal-Bench Hard2.27Oct 8, 2026aa_terminalbench_hard
0 of 7 ranked benchmarks measured
- 0.00Oct 8, 2026
- τ-Bench V3 · Banking5.77Aug 10, 2026tauBanking
0 of 3 ranked benchmarks measured
- 27.33Oct 8, 2026
0 of 4 ranked benchmarks measured
- 6.70Jun 19, 2026
- AA-Omniscience · Accuracy13.20Oct 8, 2026omniscienceAccuracy
- AA-Omniscience · Non-hallucination58.85Oct 8, 2026omniscienceNonHallucination
More results
Benchmarks outside the capability baskets. They are not ranked against a leader.
- livebench_language39.50Aug 23, 2026livebench_language@2025-04-07
- 85.40Jul 2, 2026
- 81.50Jul 2, 2026
- 28.00Jun 19, 2026
- livebench_language39.50Jun 17, 2026livebench_language@2025-04-07
- 34.39May 10, 2026
Show 19 more resultsHide 19 results
- 1.92Aug 10, 2026
- AA Intelligence8.94Oct 8, 2026aa_intelligence_index
- -22.52Oct 8, 2026
- Artificial Analysis Coding Index15.89Sep 9, 2026aa_coding_index
- 83.10Oct 7, 2026
- 88.10Oct 7, 2026
- 88.10Aug 25, 2026
- 87.17May 10, 2026
- 69.20Aug 25, 2026
- 85.60Aug 25, 2026
- 85.60Oct 7, 2026
- 67.13May 10, 2026
- 77.60Aug 25, 2026
- 65.00Oct 7, 2026
- 65.00Aug 25, 2026
- 76.29May 10, 2026
- TAU-bench (airline)22.80Oct 7, 2026TAU-bench Airline
- TAU-bench (retail)51.00Oct 7, 2026TAU-bench Retail
- τ²-Bench Telecom (AA run)24.56Oct 8, 2026aa_tau2
Claude 3.5 Haiku: common questions
Who makes Claude 3.5 Haiku?
Claude 3.5 Haiku is made by Anthropic.
When was Claude 3.5 Haiku released?
Claude 3.5 Haiku was released on Oct 22, 2024, according to Artificial Analysis.
What is Claude 3.5 Haiku good at?
Claude 3.5 Haiku is behind the leaders in multimodal tasks and instruction following. Too few results yet to rate reasoning, coding, agentic tasks, safety, long context, math, multilingual tasks, or factuality.
How much does Claude 3.5 Haiku cost?
Claude 3.5 Haiku costs $0.80 per million input tokens and $4.00 per million output tokens, according to Anthropic's own price page. At a mix of three input tokens to one output token, it costs more than 68% of the 331 priced models we track.
How many benchmarks has Claude 3.5 Haiku been tested on?
We track 46 results for Claude 3.5 Haiku on 37 benchmarks from 12 sources, 14 of them independently verified. The latest was recorded on Oct 8, 2026.
Which API features does Claude 3.5 Haiku support?
OpenRouter lists tool calling for Claude 3.5 Haiku.
About this record
Where Claude 3.5 Haiku's numbers come from, and every name it appears under.
- Tracked since
- Apr 25, 2026
- Newest source mention
- Sep 24, 2026
Where the results come from
Verification: 46 scores · 14 independently verified · 18 aggregator-attributed · 6 vendor cross-reference · 8 vendor-reported. How these tiers are assigned
From 12 sources on 6 sites. Artificial Analysis supplies 18 of them; the 14 independently verified results come from 4 sites. Bars are coloured by trust tier.
- artificialanalysis.ai18
- huggingface.co12
- api.llm-stats.com8
- storage.googleapis.com6
- aider.chat1
- epoch.ai1