Compare · ModelsLive · 2 picked · head to head
Kimi K2.6 vs Qwen3.6 35B A3B
Side by side · benchmarks, pricing, and signals you can act on.
Winner summary
Kimi K2.6 wins on 9/11 benchmarks
Kimi K2.6 wins 9 of 11 shared benchmarks. Leads in speed · general · math.
Category leads
speed·Kimi K2.6knowledge·Qwen3.6 35B A3Bgeneral·Kimi K2.6math·Kimi K2.6coding·Kimi K2.6
Hype vs Reality
Attention vs performance
Kimi K2.6
#211 by perf·#18 by attention
Qwen3.6 35B A3B
#181 by perf·#2 by attention
Best value
Qwen3.6 35B A3B
4.7x better value than Kimi K2.6
Kimi K2.6
15.4 pts/$
$2.48/M
Qwen3.6 35B A3B
72.7 pts/$
$0.57/M
Vendor risk
Who is behind the model
Moonshot AI
$50.0B·Tier 1
Alibaba (Qwen)
$293.0B·Tier 1
Head to head
11 benchmarks · 2 models
Kimi K2.6Qwen3.6 35B A3B
Artificial Analysis · Agentic Index
Kimi K2.6 leads by +8.9
Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?"
Kimi K2.6
30.3
Qwen3.6 35B A3B
21.4
Artificial Analysis · Coding Index
Kimi K2.6 leads by +14.1
Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads.
Kimi K2.6
56.0
Qwen3.6 35B A3B
41.9
Artificial Analysis · Quality Index
Kimi K2.6 leads by +24.6
Kimi K2.6
42.8
Qwen3.6 35B A3B
18.2
Chess Puzzles
Qwen3.6 35B A3B leads by +6.3
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Kimi K2.6
15.8
Qwen3.6 35B A3B
22.1
Dtbench
Kimi K2.6 leads by +28.4
Kimi K2.6
84.9
Qwen3.6 35B A3B
56.5
FrontierMath-Tiers-1-3-v2-Private
Kimi K2.6 leads by +36.8
Kimi K2.6
57.2
Qwen3.6 35B A3B
20.4
GPQA diamond
Kimi K2.6 leads by +7.9
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Kimi K2.6
87.7
Qwen3.6 35B A3B
79.8
Lmca
Kimi K2.6 leads by +8.9
Kimi K2.6
43.8
Qwen3.6 35B A3B
34.9
Mystery Game Puzzles
Qwen3.6 35B A3B leads by +4.4
Kimi K2.6
9.7
Qwen3.6 35B A3B
14.1
OTIS Mock AIME 2024-2025
Kimi K2.6 leads by +9.5
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Kimi K2.6
96.1
Qwen3.6 35B A3B
86.7
WeirdML
Kimi K2.6 leads by +21.4
WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns.
Kimi K2.6
55.9
Qwen3.6 35B A3B
34.5
Full benchmark table
| Benchmark | Kimi K2.6 | Qwen3.6 35B A3B |
|---|---|---|
Artificial Analysis · Agentic Index Artificial Analysis Agentic Index · a composite score measuring how well a model performs in agentic workflows · multi-step tool use, planning, error recovery, and autonomous task completion. Aggregates results from multiple agentic benchmarks including SWE-bench, tool-use tests, and planning evaluations. The canonical single-number metric for "how good is this model as an agent?" | 30.3 | 21.4 |
Artificial Analysis · Coding Index Artificial Analysis Coding Index · a composite score that aggregates performance across multiple coding benchmarks into a single index. Tracks code generation quality, debugging ability, multi-language competence, and real-world software engineering tasks. Used by Artificial Analysis to rank model coding capability in a normalized, comparable format. Useful for developers choosing between models for coding-heavy workloads. | 56.0 | 41.9 |
Artificial Analysis · Quality Index | 42.8 | 18.2 |
Chess Puzzles Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities. | 15.8 | 22.1 |
Dtbench | 84.9 | 56.5 |
FrontierMath-Tiers-1-3-v2-Private | 57.2 | 20.4 |
GPQA diamond Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs. | 87.7 | 79.8 |
Lmca | 43.8 | 34.9 |
Mystery Game Puzzles | 9.7 | 14.1 |
OTIS Mock AIME 2024-2025 OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills. | 96.1 | 86.7 |
WeirdML WeirdML · tests models on unusual and adversarial machine learning tasks that require creative problem-solving beyond standard patterns. | 55.9 | 34.5 |
Pricing · per 1M tokens · projected $/mo at 10M tokens
| Model | Input | Output | Context | Projected $/mo |
|---|---|---|---|---|
| $0.95 | $4.00 | 262K tokens (~131 books) | $17.13 | |
| $0.15 | $1.00 | 262K tokens (~131 books) | $3.63 |