Anthropic

Claude 3.5 Haiku

Type
language model
Context
200K tokens
Max output
8K
Released
22 October 2024
Knowledge cutoff
July 2024
API string
claude-3-5-haiku-20241022

Prices by provider

Provider Input, $ per 1M Output, $ per 1M In our data since
Amazon Bedrock eu.anthropic.claude-3-5-haiku-20241022-v1:0 $0.25 $1.25 2 Aug 2026
302.AI $0.8 $4 1 Aug 2026
Amazon Bedrock anthropic.claude-3-5-haiku-20241022-v1:0 $0.8 $4 2 Aug 2026
NanoGPT $0.8 $4 1 Aug 2026

Only the base tier and only per-token prices are shown. Batch, discounted and cached rates, as well as prices per image or per second of video, do not go into this table: they cannot stand in the same column as a price per million tokens.

A single provider appears several times if it sells this model under different identifiers and at different prices — for example, at different weight precision. The identifier is shown next to the name. Identical offers that arrived from two sources under different spellings are merged into one row.

The date in the last column is the day this price first entered our collection. The price may well be older: before that day we simply were not recording it. It has not changed since — otherwise a new row with a new date would stand in its place.

Measurement results

Task set Result Run conditions Measured by
Arena Score, programming 1,286.65 1,281–1,293
Arena Score in English 1,268.85 1,265–1,273
Arena Score, multi-turn dialogue 1,264.76 1,259–1,271
Arena Score in French 1,261.9 1,239–1,285
Arena Score, long queries 1,260.29 1,254–1,266
Arena Score, overall 1,255.44 1,252–1,259
Arena Score in Russian 1,252.79 1,245–1,260
Arena Score in Spanish 1,251.6 1,231–1,272
Arena Score, hard prompts 1,251 1,246–1,256
Arena Score, mathematics 1,244.48 1,237–1,252
Arena Score, instruction following 1,240.64 1,236–1,245
Arena Score, creative writing 1,232.59 1,226–1,239
Arena Score, expert questions 1,207.2 1,197–1,218
Arena Score, working with images 1,092.78 1,077–1,109
Arena Score, text recognition in images 1,089.62 1,070–1,109
Arena Score, understanding diagrams 1,079.26 1,047–1,112
Creative writing (Lech Mazur’s evaluation) 73.5 % lechmazur/writing Github repository
MATH, difficulty level five 46.36 % · with a tuned harness Epoch evaluations
GeoBench — locating a place from a photograph 34 % GeoBench leaderboard
CadEval — building CAD models with code 32 % CadEval Dashboard
WeirdML — unusual machine learning tasks 30.73 % WeirdML Leaderboard
Aider Polyglot — code edits in six languages 28 % · with a tuned harness Aider LLM Leaderboards
BALROG — game environments 19.3 % Balrog Leaderboard
GPQA Diamond — graduate-level questions 17.51 % · with a tuned harness Epoch evaluations
SimpleQA Verified — factual accuracy 6.7 % Epoch evaluations
Mock AIME 2024–2025 — olympiad problems 4.21 % · with a tuned harness Epoch evaluations
CritPt — physics problems 0 %