GenMagic logoGenMagic
Sign in

Claude Haiku 5.5 vs DeepSeek V4 Flash 0731

Anthropic against DeepSeek. Prices, scores and speed below are live.

  • DeepSeek V4 Flash 0731 costs 40% less on blended price ($0.18 against $0.30 per 1M tokens, 3 input to 1 output).
  • Claude Haiku 5.5 scores higher on the Artificial Analysis Intelligence Index: 36.5 against 34.5.
  • On the 26 benchmarks both have scores on, Claude Haiku 5.5 leads on 18 and DeepSeek V4 Flash 0731 on 7, with 1 tied.
  • DeepSeek V4 Flash 0731 reads more context: 1M tokens against 1M.
  • Claude Haiku 5.5 is newer: listed Oct 7, 2026, against Jul 31, 2026.
Run both on your promptClaude Haiku 5.5DeepSeek V4 Flash 0731

Side by side

Claude Haiku 5.5DeepSeek V4 Flash 0731
LabAnthropicDeepSeek
ListedOct 7, 2026Jul 31, 2026
Input per 1M tokens$0.15$0.088
Output per 1M tokens$0.75$0.45
Cached input per 1M$0.015$0.052
Blended (3:1) per 1M$0.30$0.18
Context window1M1M
Max output128K944K
Knowledge cutoff--
Intelligence index36.534.5
Coding index-69.1
ReasoningYesYes
Reads imagesYesNo
Tool useYesYes
Structured outputYesYes

Every benchmark both are scored on

Each model's best score per benchmark, with its source. Scores from different sources or settings are not always comparable; each links to where it was published.

BenchmarkClaude Haiku 5.5DeepSeek V4 Flash 0731
AIME98.9%Epoch AI94.4%Epoch AI
ARC-AGI-189%Epoch AI89%Epoch AI
ARC-AGI-261.3%Epoch AI61.4%Epoch AI
Artificial Analysis Intelligence Index36.5Artificial Analysis and OpenRouter34.5Artificial Analysis and OpenRouter
Chess Puzzles55%Epoch AI32%Epoch AI
Design Arena: 3D1337 EloDesign Arena1205 EloDesign Arena
Design Arena: ASCII Art1210 EloDesign Arena1088 EloDesign Arena
Design Arena: Code1281 EloDesign Arena1228 EloDesign Arena
Design Arena: Data Viz1290 EloDesign Arena1180 EloDesign Arena
Design Arena: Game Dev1319 EloDesign Arena1207 EloDesign Arena
Design Arena: SVG1203 EloDesign Arena1180 EloDesign Arena
Design Arena: UI Component1278 EloDesign Arena1230 EloDesign Arena
Design Arena: Website1262 EloDesign Arena1236 EloDesign Arena
DTBench95%Epoch AI90.9%Epoch AI
FrontierCode46.4Epoch AI18.8Epoch AI
FrontierMath Tier 4 (v2)46.3%Epoch AI24.4%Epoch AI
FrontierMath Tiers 1-3 (v2)75.1%Epoch AI57.5%Epoch AI
GPQA Diamond86.7%Artificial Analysis and OpenRouter86.8%Artificial Analysis and OpenRouter
LMArena (Coding)1454 EloLMArena1456 EloLMArena
LMArena (Overall)1401 EloLMArena1432 EloLMArena
LMCA51.1Epoch AI41.7Epoch AI
Mystery Game Puzzles30%Epoch AI34%Epoch AI
ProofBench89%Epoch AI56%Epoch AI
SciCode55%Epoch AI49.9%Epoch AI
SimpleQA Verified23.8%Epoch AI33.6%Epoch AI
tau-bench (Airline)72.8%Artificial Analysis and OpenRouter74%Artificial Analysis and OpenRouter

More comparisons

Where these figures come from

Prices are GenMagic's own, what the studio and the API charge, recomputed from each provider's price list every few minutes. The indices are the Artificial Analysis indices, and every benchmark score links to its source, re-read daily. Measured speed is GenMagic's own generation records, in aggregate. See Anthropic models, DeepSeek models, every LLM's price and our sources.