LIVE INTELLIGENCE · part of the Xentir AI Index

GPT-5 mini vs Kimi K2 Thinking

Name: GPT-5 mini vs Kimi K2 Thinking — benchmark results
Creator: Epoch AI
License: https://creativecommons.org/licenses/by/4.0/

GPT-5 mini · OpenAI, released 2025-08-07
Kimi K2 Thinking OPEN WEIGHTS · Moonshot, released 2025-11-06

“Epoch AI, 'AI Benchmarking Hub'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks' [online resource].” · epoch.ai/benchmarks · Licensed CC BY 4.0 · data last verified 2026-09-09 · Methodology · Data Sources

GPT-5 mini leads on more of the 3 benchmarks where both are measured — GPT-5 mini on 2, Kimi K2 Thinking on 1. The remaining 6 of 9 tracked benchmarks cannot be compared: 6 have a result for only one of the two models, 0 for neither. Nothing on this page is estimated to fill those in.

Head-to-head

Measured on 3 of 9 tracked benchmarks for both models — GPT-5 mini leads on 2, Kimi K2 Thinking leads on 1.

Benchmark	GPT-5 mini	Kimi K2 Thinking	Leads
Epoch Capabilities Index (ECI)	145.5	145.8	Kimi K2 Thinking
SWE-bench Verified	0.647 (medium)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result
GPQA Diamond	0.750 (high)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result
MATH Level 5	0.978 (high)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result
OTIS Mock AIME 2024-2025	0.867 (high)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result
FrontierMath	0.272 (high)	0.214 (moonshotai)	GPT-5 mini
FrontierMath Tier 4	0.062 (high)	0.000 (moonshotai)	GPT-5 mini
SimpleQA Verified	0.216 (high)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result
Chess Puzzles	0.300 (high)	Not measured	Comparison unavailable — Kimi K2 Thinking has no published result

Benchmark results measure specific tasks and should not be treated as a universal measure of model quality. A model with no row here for a benchmark has not been measured on it — that is left blank, never estimated.

What remains uncertain

Kimi K2 Thinking has no published result on Chess Puzzles, GPQA Diamond, MATH Level 5, OTIS Mock AIME 2024-2025, SWE-bench Verified, SimpleQA Verified.

A missing result means Xentir has not found one published by the source it cites, not that the model performs badly. See Data Sources.

Reference price

Per 1M tokens	GPT-5 mini	Kimi K2 Thinking
Output	$2.00	—

List price as published by the vendor. Xentir does not verify negotiated or enterprise rates, and price is not a measure of capability — see Data Sources.

Full profiles

GPT-5 mini → · Kimi K2 Thinking → · the full AI Model Rankings table