LLM

Claude Haiku 4.5

by Anthropic

Prices last updated November 22, 2026

Latest Haiku tier with improved capabilities at fast speed and low cost.

Context
200k
Providers
12
Cheapest Input
$0.275
Vision Function Calling JSON Mode Streaming Chat / InstructVision

Price Comparison

Compare pricing across 12 providers serving Claude Haiku 4.5

Provider Input $/1M Output $/1M Context Speed TTFT Status
KIE AI
KIE AI
$0.275 ● $1.43 — — — Live
Muapi
Muapi
$0.600 $3.00 — — — Live
PoYo
PoYo
$0.800 $4.00 — — — Live
Replicate
Replicate
$1.00 $5.00 — — — Live
OpenRouter
OpenRouter
$1.00 $5.00 — — — Live
Anthropic
Anthropic
$1.00 $5.00 — — — Live
Amazon Bedrock
Amazon Bedrock
$1.00 $5.00 200k — — Live
DeepInfra
DeepInfra
$1.00 $5.00 200k — — Live
WaveSpeed AI
WaveSpeed AI
$1.00 $5.00 200k — — Live
Infron
Infron
$1.00 $5.00 200k — — Live
Atlas Cloud
Atlas Cloud
$1.00 $5.00 200k — — Live
AIMLAPI
AIMLAPI
$1.38 $6.88 — — — Live
● Cheapest ● Fastest Prices in USD per 1M tokens

Benchmarks

Independent results for Claude Haiku 4.5, as each source publishes them. Where a source tested several versions (effort levels, thinking modes, dated snapshots), the best one is shown and named under "As tested". Scores from different sources use different scales and aren't comparable with each other.

BenchmarkScoreRankAs tested
Overall
One score combining a model's results across many benchmarks, on a common scale.
142.7±1.7
#111 of 258
Claude Haiku 4.5
Reasoning
GPQA DiamondEpoch AI
Graduate-level multiple-choice questions in biology, physics and chemistry, written to be hard to look up.
71.2%±6.3
#120 of 232
claude-haiku-4-5-20251001_32K
Oct 2025
Math
Competition maths problems from the OTIS mock AIME exams (2024-2025).
66.7%±13.9
#120 of 215
claude-haiku-4-5-20251001_32K
Oct 2025
MATH Level 5Epoch AI
The hardest level of the MATH competition-problem dataset.
96.4%±1.0
#9 of 94
claude-haiku-4-5-20251001_32K
Oct 2025
Coding
Text Arena votes on coding prompts only.
1,482±4
39,463 votes
#118 of 421
Claude Haiku 4.5 (20251001)
Nov 2026
Human votes on web apps the models build from a prompt.
1,330±5
28,713 votes
#137 of 161
Claude Haiku 4.5 (20251001)
Nov 2026
Knowledge
Short factual questions answered without search, measuring what the model knows and whether it makes things up.
13.2%±2.1
#90 of 97
claude-haiku-4-5-20251001
Aug 2026
Chat
Text ArenaLMArena
Human votes on chat answers across all kinds of prompts (style-controlled).
1,414±2
153,650 votes
#154 of 426
Claude Haiku 4.5 (20251001)
Nov 2026

Epoch AI: Epoch runs each benchmark itself on the models it evaluates, with the same harness for every model. The Epoch Capabilities Index combines many benchmarks into one score. Data: Epoch AI, 'Capabilities & benchmarking', epoch.ai/benchmarks (CC BY 4.0).

LMArena: People compare two anonymous models side by side and vote for the better answer; ratings come from those votes (Bradley-Terry, style-controlled for text and vision). Data: LMArena (arena.ai) leaderboard dataset (CC BY 4.0).

Rank is among all models the source ranks, including ones we don't list.

Claude Haiku 4.5 guides

Subscribe to our newsletter

News and updates, 1–2 emails per month. No spam.