Compare

DeepSeek V4 vs Claude Haiku 4.5

DeepSeek vs Anthropic. Specs, benchmarks, and real per-task cost — all in one page.

Verdict

DeepSeek V4 leads on LMSYS ELO (1334 vs 1305). DeepSeek V4 is ~9.1× cheaper on a 3:1 input:output blend. DeepSeek V4 is open-weights, the other is proprietary.

DeepSeek

DeepSeek V4

The shock-the-market moment of 2026. Frontier coding quality at 1/50th the price of Opus.

Unbeatable $/qualityOpen weightsGreat at code

Anthropic

Claude Haiku 4.5

Fast, cheap, and smart enough for most routing and extraction tasks. Great base model for subagents.

Sub-second latency10x cheaper than OpusStrong structured output

Pricing

Input / 1M$0.14$0.80

Output / 1M$0.28$4.00

Context128K200K

Max output16K16K

LicenseOpen weightsProprietary

Released2026-02-202025-10-01

Benchmarks

LMSYS ELO1334.01305.0

MMLU Pro85.481.4

HumanEval91.587.0

SWE-Bench54.149.2

MATH88.776.8

IFEval85.887.3

Per-task cost

Summarize a 1-hour meeting transcript$0.0022$0.014

Review a 500-line pull request$0.0017$0.014

Answer a customer support ticket$0.0006$0.0056

Extract structured data from a resume$0.0010$0.0080

Debug a stack trace with context$0.0014$0.013

Per-call cost using published token counts for each task. Real-world prompts vary.

Tool

Interactive compare →

Add up to four models, tweak tasks live.