VECTOR WIREAI INTELLIGENCE
UTC
Refresh Models Deals Regulatory Sources

DeepSeek-V4-Pro vs MiniMax M2.7

74 SHARED BENCHMARKS

Across 74 shared benchmarks, DeepSeek-V4-Pro scores higher on 63 and MiniMax M2.7 on 10, with 1 level. The widest gap is OmniScience Non-Hallucination, where MiniMax M2.7 scores 65.6 against 12.2. MiniMax M2.7 is the cheaper of the two on tracked API pricing ($0.30 against $1.32 per million input tokens).

DEEPSEEKVSMINIMAX74 SHARED6310 HEAD-TO-HEAD
AA-LCR7075.3
aime_202696.789.8
Airline80.875.3
apexAgents24.310.6
Banking25.914.6
browsecomp83.454.1
critpt130.6
gdpval54.647.6
GDPval-AA v213071159
GPQA Diamond90.587.4
HLE48.229.6
IFBench76.575.7
itbenchSre38.326.5
MCP Atlas74.248.8
MMLU-Pro87.581.9
nl2repo38.539.8
PinchBench88.677.6
Retail88.984.9
strongreject98.699.4
Telecom96.389.6
toolathlon51.846.3
vectara_hallucination_rate↓ lower is better8.612.9
τ³-Bench25.967.6

Best tracked score per model per benchmark (default configuration; source-attributed). ↓ marks lower-is-better metrics. Open either model for its full surface, provenance and pricing. Quoted rates are the price-setter row we currently track for each model — its direct or vendor-official listing where one exists (deepseek-official, direct), otherwise the lowest tracked offer.

Compare DeepSeek-V4-Pro withEVERY TRACKED PAIRING

Anthropic12 PAIRINGS
OpenAI11 PAIRINGS
Google11 PAIRINGS
Alibaba11 PAIRINGS
DeepSeek4 PAIRINGS
Moonshot4 PAIRINGS
Meta2 PAIRINGS
Z.ai2 PAIRINGS
NVIDIA1 PAIRING

Compare MiniMax M2.7 withEVERY TRACKED PAIRING

Anthropic12 PAIRINGS
OpenAI11 PAIRINGS
Google11 PAIRINGS
Alibaba11 PAIRINGS
DeepSeek4 PAIRINGS
Moonshot4 PAIRINGS
Meta2 PAIRINGS
Z.ai2 PAIRINGS
NVIDIA1 PAIRING