Kimi K2 Thinking vs Step 3.7 Flash
Kimi K2 Thinking leads Step 3.7 Flash on the Artificial Analysis Intelligence Index, 22.0 to 19.5, and Step 3.7 Flash costs 2.2x less per output token ($1.15 vs $2.50 per million). Updated October 21, 2026; 9 benchmarks have a result for both.
Kimi K2 Thinking or Step 3.7 Flash: which to pick
Neither model has a coding score yet, so the two cannot be ranked on coding.
Pick Kimi K2 Thinking for work measured by Terminal-Bench Hard (Artificial Analysis run) or ALE-Bench, where it leads.
- Shared results
- 9
- Kimi K2 Thinking only
- 21
- Step 3.7 Flash only
- 9
- Like-for-like categories
- 6 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
No call
Neither has a coding score yet.
One side not measured
Intelligence
AA Intelligence Index
Kimi K2 Thinking
Kimi K2 Thinking leads, 22.0 against 19.5.
Same board, both measured
Reasoning
GPQA Diamond
No call
Only Kimi K2 Thinking has a GPQA Diamond result, 78.5%.
Directional only
Math
AIME
No call
Only Kimi K2 Thinking has a AIME result, 85.4%.
Directional only
Knowledge
MMLU Pro
No call
Only Kimi K2 Thinking has a MMLU Pro result, 81.1%.
Directional only
Web design
Design Arena: websites
Step 3.7 Flash
Step 3.7 Flash leads, 1196 against 1113.
Same board, both measured
Long documents
Listed context window
No call
Both take 262K tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Step 3.7 Flash
Kimi K2 Thinking $0.0019, Step 3.7 Flash $0.0008 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Kimi K2 Thinking has the number, $0.61.
Directional only
At a glance
- Coding index
- Kimi K2 Thinking not benchmarked, Step 3.7 Flash not benchmarked.
- Widest category
- Web design, 10.4 points apart on average over 1 shared benchmark.
- Widest single gap
- Design Arena: websites: Kimi K2 Thinking 1113 against Step 3.7 Flash 1196, 10.4 points apart on a 0 to 100 scale.
- Output price
- Kimi K2 Thinking $2.50 against Step 3.7 Flash $1.15 per million output tokens, 2.2x apart.
- Context window
- Both take 262K tokens.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Kimi K2 Thinking | Bars | Step 3.7 Flash | Ties |
|---|---|---|---|---|
| Coding | 0 | 2 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 0 | 1 | ||
| Reasoning | 2 | 0 | ||
| Web design | 0 | 1 | ||
| General | 1 | 1 | ||
| All shared boards | 4 | 5 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index02
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| SWE-bench Verified | 63.4% | Not measured | Only Kimi K2 Thinking measured |
| AA Coding Index | Not measured | 39.6 | Only Step 3.7 Flash measured |
| LiveCodeBench (Vals AI) | 63.1% | Not measured | Only Kimi K2 Thinking measured |
| SWE-bench Verified (Vals AI) | 60.2% | Not measured | Only Kimi K2 Thinking measured |
| LiveCodeBench (Artificial Analysis run) | 85.3% | Not measured | Only Kimi K2 Thinking measured |
| SciCode | Not measured | 43.9% | Only Step 3.7 Flash measured |
| Terminal-Bench Hard (Artificial Analysis run) | 31.1% | 35.6% | Step 3.7 Flash |
| Terminal-Bench | 35.7% | Not measured | Only Kimi K2 Thinking measured |
| WeirdML | 42.8% | Not measured | Only Kimi K2 Thinking measured |
| ALE-Bench | 597.5 | 694.1 | Step 3.7 Flash |
| Neither model has a result on the other 28 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| AA Intelligence Index | 22.0 | 19.5 | Kimi K2 Thinking |
| Epoch Capabilities Index | 145.8 | Not measured | Only Kimi K2 Thinking measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| τ²-Bench Telecom | 93.0% | 98.5% | Step 3.7 Flash |
| METR time horizons | 59.2% | Not measured | Only Kimi K2 Thinking measured |
| Neither model has a result on the other 37 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| GPQA Diamond | 78.5% | Not measured | Only Kimi K2 Thinking measured |
| GPQA Diamond (Artificial Analysis run) | 83.8% | 80.9% | Kimi K2 Thinking |
| Humanity's Last Exam (Artificial Analysis run) | 23.8% | 21.4% | Kimi K2 Thinking |
| Neither model has a result on the other 18 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index4 boards, none shared
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| AIME | 85.4% | Not measured | Only Kimi K2 Thinking measured |
| MGSM (Vals AI) | 90.2% | Not measured | Only Kimi K2 Thinking measured |
| AA Math Index | 94.7 | Not measured | Only Kimi K2 Thinking measured |
| AIME 2025 (Artificial Analysis run) | 94.7% | Not measured | Only Kimi K2 Thinking measured |
| Neither model has a result on the other 9 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| MMLU Pro | 81.1% | Not measured | Only Kimi K2 Thinking measured |
| LegalBench (Vals AI) | 80.2% | Not measured | Only Kimi K2 Thinking measured |
| CaseLaw (Vals AI) | 65.7% | Not measured | Only Kimi K2 Thinking measured |
| CorpFin (Vals AI) | 60.6% | Not measured | Only Kimi K2 Thinking measured |
| TaxEval (Vals AI) | 71.7% | Not measured | Only Kimi K2 Thinking measured |
| MedQA (Vals AI) | 92.6% | Not measured | Only Kimi K2 Thinking measured |
| MMLU Pro (Artificial Analysis run) | 84.8% | Not measured | Only Kimi K2 Thinking measured |
| Neither model has a result on the other 10 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| Design Arena: websites | 1113 | 1196 | Step 3.7 Flash |
| Design Arena: UI components | Not measured | 1179 | Only Step 3.7 Flash measured |
| Design Arena: game dev | Not measured | 1158 | Only Step 3.7 Flash measured |
| Design Arena: data viz | Not measured | 1172 | Only Step 3.7 Flash measured |
| Design Arena: 3D | Not measured | 1129 | Only Step 3.7 Flash measured |
| Design Arena: all code categories | Not measured | 1182 | Only Step 3.7 Flash measured |
| Design Arena: SVG | Not measured | 1085 | Only Step 3.7 Flash measured |
| Design Arena: ASCII art | Not measured | 1152 | Only Step 3.7 Flash measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index11
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| IFBench | 68.1% | 67.3% | Kimi K2 Thinking |
| Long context reasoning (AA-LCR) | 72.0% | 73.7% | Step 3.7 Flash |
| CL-bench | 17.6% | Not measured | Only Kimi K2 Thinking measured |
| Neither model has a result on the other 43 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Kimi K2 Thinking | Step 3.7 Flash | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Step 3.7 Flash costs less
Repository review
50K input + 3K output tokens
Step 3.7 Flash costs less
Agent loop
200K input + 10K output tokens
Step 3.7 Flash costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | MoonshotAI Kimi K2 Thinking | StepFun Step 3.7 Flash |
|---|---|---|
| Maker | Moonshot AI | StepFun |
| Weights | Open weights | Open weights |
| Released | November 6, 2025 | May 28, 2026 |
| Context window | 262K | 262K |
| Input $/Mtok | $0.60 | $0.20 |
| Output $/Mtok | $2.50 | $1.15 |
| Coding index | Not benchmarked | Not benchmarked |
| Independent measurements behind it | 1 | 0 |
| Cost per solved task | $0.61 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Kimi K2 Thinking better than Step 3.7 Flash for coding?
Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Kimi K2 Thinking for work measured by Terminal-Bench Hard (Artificial Analysis run) or ALE-Bench, where it leads.
Which is cheaper, Kimi K2 Thinking or Step 3.7 Flash?
Kimi K2 Thinking costs $0.60 per million input tokens and $2.50 per million output tokens; Step 3.7 Flash costs $0.20 and $1.15. Cost per solved task is known for Kimi K2 Thinking only.
Which has the bigger context window, Kimi K2 Thinking or Step 3.7 Flash?
Neither: both take 262K tokens.
How many benchmarks is the Kimi K2 Thinking vs Step 3.7 Flash comparison based on?
9 benchmarks have a result for both models as of October 21, 2026. Kimi K2 Thinking's coding index rests on 1 independent benchmark and Step 3.7 Flash's on 0. Where only one model has a result, the page says so instead of naming a winner.