DeepSeek Coder V2 Lite Instruct — benchmark results

DeepSeek's 16B MoE coder (2.4B active) with a 128K context, supporting 338 programming languages. Provider: DeepSeek. Released 2024-06-17. Access: Open.

Unified ELO 1438 ± 29, rank #1051 of 1776 rated models, from 30 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
GSM8K87.6Accuracy (%)88.3
CanAiCode100Junior-v2 Python Pass Rate (%)85.6
LiveBench AMPS Hard33Score75
LiveBench Math Comp33.33Score73.6
LiveBench Table Reformat52Score62.5
LiveBench Coding Completion23.68Score59
LiveBench Web Of Lies V224Score54.2
LiveBench LCB Generation30Score53.5
LiveBench Table Join9Score50
LiveBench Olympiad36.98Score44.4
LiveBench Typos20Score44.4
BigCodeBench36.8Pass@1 (%)39.3

Interactive version: theaggregate.ai/model?slug=deepseek-coder-v2-lite-instruct · How the rankings work · Data refreshed daily, snapshot 2026-07-22.