DeepSeek V4.1-Flash-1017 vs Claude Haiku 5.5: more accurate at a fraction of the price

The New Stack benchmarked DeepSeek's V4.1-Flash-1017 update against Anthropic's Claude Haiku 5.5. DeepSeek scored 6/6 on an agentic bug fix and 1600/1600 triage fields versus 5/6 and 1596 for Haiku, while both tied at 20/20 on SQL queries; total cost was $0.07 versus $0.30.

The New Stack

Open source

$DeepSeek V4.1-Flash-1017 vs Claude Haiku 5.5: more accurate at a fraction of the price$

DeepSeek scored 6/6 on the agentic bug fix, Haiku 5.5 scored 5/6
On 400 triage reports DeepSeek filled 1600 fields correctly, Haiku 1596
Both models answered all 20 SQL questions correctly
Three-test cost: $0.07 for DeepSeek versus $0.30 for Haiku

#DeepSeek#Anthropic#Claude

DeepSeek V4.1-Flash-1017 vs Claude Haiku 5.5: more accurate at a fraction of the price

Haiku 5.5 vs 4.5: 100-prompt test shows far less repeatable answers

Anthropic launches Claude Haiku 5.5 with 1M context and up to 90% price cuts

Test: Gemini 4 Argon is half the price of Claude Opus 5.5 but far less accurate

Mistral Large 4 and Claude Haiku 5.5 headline new AI model releases