// COMPARISON

Qwen3-235B-A22B-Thinking (Jul 2025) vs Kimi K2 Thinking

A head-to-head benchmark comparison of Qwen3-235B-A22B-Thinking (Jul 2025) vs Kimi K2 Thinking across 48 evaluations. Kimi K2 Thinking leads on the most benchmarks (27).

The models

Leads 27
MO

Kimi K2 Thinking

Moonshot
Language Open (Restricted)
Params
1T
Context
β€”
Released
Nov 2025

Today, we are introducing Kimi K2 Thinking, our best open-source thinking model. Built as a thinking agent, it reasons step by step while using tools, achieving.

Full Kimi K2 Thinking specs β†’

Benchmark comparison

Benchmark Qwen3-235B-A22B-Thinking (Jul 2025) Kimi K2 Thinking
agentic
BrowseComp β€” 60.2%
BrowseComp-Zh β€” 62.3
FinSearchComp-T3 β€” 47.4
MCP-Mark β€” 20.4
MCP-Universe β€” 35.6
Seal-0 β€” 56.3
Terminal-Bench 2.0 β€” 47.1%
Tool-Decathlon β€” 17.6
τ²-Bench (TAU2) β€” 74.3
code
Codeforces Elo 2,055 β€”
LiveCodeBench 78.2% 82.6%
LiveCodeBench v5 70.7% β€”
LiveCodeBench v6 75.72% 83.1%
SWE-bench Verified β€” 71.3%
SciCode 42.9% 44.8%
coding
Aider Benchmark 88.91 β€”
Multi-SWE-bench β€” 41.9
OJ-Bench (cpp) β€” 48.7
SWE-Bench Multilingual β€” 61.1
general
ArenaHard v2 80.18 β€”
C-Eval 93.13% β€”
IFEval 87.8% β€”
Longform Writing β€” 73.8
MMLU-Pro 84.5% 84.6%
MMLU-Redux 92.7% 94.4%
MultiChallenge 52.75% β€”
knowledge
AGIEval 90.01 β€”
CMMLU 90.64 β€”
TriviaQA 79.63 β€”
long-context
Frames β€” 87
math
AIME 2024 94.1% β€”
AIME 2025 92.3% 100%
CNMO 2024 89.5% β€”
HMMT February 2025 83.9% 97.5%
HMMT November 2025 β€” 89.2
IMOAnswerBench β€” 78.6
MATH-500 98% β€”
Omni-MATH 82.52 β€”
UGMathBench 77 β€”
reasoning
ARC-AGI-1 48.12 β€”
BBEH 60 β€”
FinanceReasoning 87.65 β€”
GPQA Diamond 81.1% 84.5%
Humanity's Last Exam 15.8% 51%
ZebraLogic 97.03% β€”
safety
HealthBench 55.56% 58%
science
PhyBench 42.61 β€”
tool-use
BFCL v3 73.53% β€”

Best result per row highlighted in cyan. Each benchmark links to its definition and sources; each model links to its full scorecard.

Frequently asked questions

Which model is best, Qwen3-235B-A22B-Thinking (Jul 2025) vs Kimi K2 Thinking? β–Ά

Across the 48 benchmarks compared here, Kimi K2 Thinking leads on the most (27). The right choice still depends on your task β€” check the per-benchmark table above for coding, reasoning and math.

Which of these models is best for coding? β–Ά

Qwen3-235B-A22B-Thinking (Jul 2025) has the strongest result on Codeforces Elo among the models compared here. See the code-category rows in the table for the full picture.

How many benchmarks are compared? β–Ά

This comparison covers 48 benchmarks on which at least one of the selected models has a published score.

Where do the benchmark scores come from? β–Ά

Scores are aggregated from official model cards, technical reports and standard public evaluations, and link back to each benchmark's source. They are updated as new results are published.

Explore further

β†’ Qwen3-235B-A22B-Thinking (Jul 2025) β†’ Kimi K2 Thinking β†’ AI benchmarks leaderboard β†’ Build a different comparison
Build a different comparison

Models (0 selected)

Benchmarks (auto)

Select at least two models.