Kimi K2.7 Code vs Llama 3.2 3B Instruct
Kimi K2.7 Code leads Llama 3.2 3B Instruct on the Artificial Analysis Intelligence Index, 25.8 to 5.7, and Llama 3.2 3B Instruct costs 12.0x less per output token ($0.3325 vs $4.00 per million). Updated October 21, 2026; 6 benchmarks have a result for both.
Kimi K2.7 Code or Llama 3.2 3B Instruct: which to pick
Only Kimi K2.7 Code has a coding score, 53.4 on the Vibe Coding Index; Llama 3.2 3B Instruct has none, so the two cannot be ranked on coding.
Pick Llama 3.2 3B Instruct for output tokens 12.0x cheaper ($0.3325 vs $4.00 per million).
- Shared results
- 6
- Kimi K2.7 Code only
- 45
- Llama 3.2 3B Instruct only
- 29
- Like-for-like categories
- 4 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
Only Kimi K2.7 Code has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
Kimi K2.7 Code
Kimi K2.7 Code leads, 25.8 against 5.7.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Kimi K2.7 Code has a AA Agentic Index result, 21.0.
Directional only
Web design
Design Arena: websites
No call
Only Kimi K2.7 Code has a Design Arena: websites result, 1264.
Directional only
Long documents
Listed context window
Kimi K2.7 Code
Kimi K2.7 Code takes 262K tokens against 131K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Llama 3.2 3B Instruct
Kimi K2.7 Code $0.0029, Llama 3.2 3B Instruct $0.0002 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Kimi K2.7 Code has the number, $0.35.
Directional only
At a glance
- Coding index
- Kimi K2.7 Code 53.4 from 4 benchmarks, Llama 3.2 3B Instruct not benchmarked.
- Widest category
- Agentic work, 69.0 points apart on average over 1 shared benchmark.
- Widest single gap
- Long context reasoning (AA-LCR): Kimi K2.7 Code 79.3% against Llama 3.2 3B Instruct 4.3%, 75.0 points apart on a 0 to 100 scale.
- Output price
- Kimi K2.7 Code $4.00 against Llama 3.2 3B Instruct $0.3325 per million output tokens, 12.0x apart.
- Context window
- Kimi K2.7 Code takes 262K tokens, Llama 3.2 3B Instruct 131K.
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 K2.7 Code | Bars | Llama 3.2 3B Instruct | Ties |
|---|---|---|---|---|
| Intelligence | 1 | 0 | ||
| Agentic work | 1 | 0 | ||
| Reasoning | 2 | 0 | ||
| General | 2 | 0 | ||
| All shared boards | 6 | 0 |
CodingDirectional onlySix of these boards feed the Vibe Coding Index20 boards, none shared
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| LMArena WebDev | 1474 | Not measured | Only Kimi K2.7 Code measured |
| AA Coding Index | 60.8 | Not measured | Only Kimi K2.7 Code measured |
| LiveCodeBench (Vals AI) | 82.0% | Not measured | Only Kimi K2.7 Code measured |
| SWE-bench Verified (Vals AI) | 78.2% | Not measured | Only Kimi K2.7 Code measured |
| Vibe Code Bench (Vals AI) | 47.2% | Not measured | Only Kimi K2.7 Code measured |
| Code migration (Vals AI) | 25.4% | Not measured | Only Kimi K2.7 Code measured |
| ProgramBench (Vals AI) | 0.0% | Not measured | Only Kimi K2.7 Code measured |
| Terminal-Bench 2.1 (Vals AI) | 67.0% | Not measured | Only Kimi K2.7 Code measured |
| SkillsBench (Vals AI) | 50.0% | Not measured | Only Kimi K2.7 Code measured |
| LiveCodeBench (Artificial Analysis run) | Not measured | 8.3% | Only Llama 3.2 3B Instruct measured |
| SciCode | 47.8% | Not measured | Only Kimi K2.7 Code measured |
| Terminal-Bench Hard (Artificial Analysis run) | 44.7% | Not measured | Only Kimi K2.7 Code measured |
| FrontierCode | 30.1% | Not measured | Only Kimi K2.7 Code measured |
| WeirdML | 54.1% | Not measured | Only Kimi K2.7 Code measured |
| ALE-Bench | 886.2 | Not measured | Only Kimi K2.7 Code measured |
| LMArena Text: coding | Not measured | 1098 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: software and IT | Not measured | 1103 | Only Llama 3.2 3B Instruct measured |
| LMArena WebDev: React | 1475 | Not measured | Only Kimi K2.7 Code measured |
| LMArena WebDev: HTML | 1451 | Not measured | Only Kimi K2.7 Code measured |
| LMArena WebDev: from an image | 1513 | Not measured | Only Kimi K2.7 Code measured |
| Neither model has a result on the other 19 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| AA Intelligence Index | 25.8 | 5.7 | Kimi K2.7 Code |
| Epoch Capabilities Index | 150.1 | Not measured | Only Kimi K2.7 Code measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| AA Agentic Index | 21.0 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building web apps | 1176 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building full-stack apps | 1175 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building mobile apps | 1143 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building Android apps | 1137 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building Godot games | 1147 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: agents building games | 1131 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: HTML slides | 1201 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: PowerPoint slides | 1154 | Not measured | Only Kimi K2.7 Code measured |
| τ²-Bench Telecom | 90.1% | 21.1% | Kimi K2.7 Code |
| APEX-Agents | 37.6% | Not measured | Only Kimi K2.7 Code measured |
| BALROG | Not measured | 10.1% | Only Llama 3.2 3B Instruct measured |
| Vending-Bench 2 | $5,083 | Not measured | Only Kimi K2.7 Code measured |
| Berkeley Function Calling Leaderboard (BFCL) | Not measured | 21.9% | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 27 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| GPQA Diamond (Artificial Analysis run) | 89.6% | 25.5% | Kimi K2.7 Code |
| Humanity's Last Exam (Artificial Analysis run) | 35.0% | 5.3% | Kimi K2.7 Code |
| GPQA Diamond (Epoch AI run) | 87.9% | Not measured | Only Kimi K2.7 Code measured |
| SimpleBench | 57.9% | Not measured | Only Kimi K2.7 Code measured |
| Chess puzzles | 14.0% | Not measured | Only Kimi K2.7 Code measured |
| LMArena Text: hard prompts | Not measured | 1095 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: expert | Not measured | 1090 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: hard prompts in English | Not measured | 1141 | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 14 boards here | |||
MathDirectional onlyPublic boards, shown as published, kept out of the index7 boards, none shared
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| AA Math Index | Not measured | 3.3 | Only Llama 3.2 3B Instruct measured |
| AIME 2025 (Artificial Analysis run) | Not measured | 3.3% | Only Llama 3.2 3B Instruct measured |
| OTIS Mock AIME | 95.6% | Not measured | Only Kimi K2.7 Code measured |
| FrontierMath, tiers 1 to 3 | 54.0% | Not measured | Only Kimi K2.7 Code measured |
| FrontierMath, tier 4 | 12.2% | Not measured | Only Kimi K2.7 Code measured |
| LMArena Text: math | Not measured | 1126 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: mathematical work | Not measured | 1128 | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 6 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index6 boards, none shared
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| MMLU Pro (Artificial Analysis run) | Not measured | 34.7% | Only Llama 3.2 3B Instruct measured |
| SimpleQA Verified | 36.5% | Not measured | Only Kimi K2.7 Code measured |
| LMArena Text: science | Not measured | 1106 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: medicine and healthcare | Not measured | 1104 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: legal and government | Not measured | 1115 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: business and finance | Not measured | 1074 | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 12 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| Design Arena: websites | 1264 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: UI components | 1259 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: game dev | 1220 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: data viz | 1227 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: 3D | 1241 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: all code categories | 1253 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: SVG | 1176 | Not measured | Only Kimi K2.7 Code measured |
| Design Arena: ASCII art | 1196 | Not measured | Only Kimi K2.7 Code measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index20
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | Leader |
|---|---|---|---|
| IFBench | 63.1% | 26.2% | Kimi K2.7 Code |
| Long context reasoning (AA-LCR) | 79.3% | 4.3% | Kimi K2.7 Code |
| Surface Evolver Bench | 48.8% | Not measured | Only Kimi K2.7 Code measured |
| LMArena Text | Not measured | 1110 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: instruction following | Not measured | 1089 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: creative writing | Not measured | 1094 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: English | Not measured | 1180 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: non-English | Not measured | 1019 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: Chinese | Not measured | 1017 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: German | Not measured | 1056 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: Russian | Not measured | 949 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: multi-turn | Not measured | 1105 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: longer queries | Not measured | 1100 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: writing and language | Not measured | 1096 | Only Llama 3.2 3B Instruct measured |
| LMArena Text: entertainment and media | Not measured | 1083 | Only Llama 3.2 3B Instruct measured |
| Neither model has a result on the other 31 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Kimi K2.7 Code | Llama 3.2 3B Instruct | 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
Llama 3.2 3B Instruct costs less
Repository review
50K input + 3K output tokens
Llama 3.2 3B Instruct costs less
Agent loop
200K input + 10K output tokens
Llama 3.2 3B Instruct 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 K2.7 Code | Meta Llama 3.2 3B Instruct |
|---|---|---|
| Maker | Moonshot AI | Meta Llama |
| Weights | Open weights | Open weights |
| Released | June 12, 2026 | September 25, 2024 |
| Context window | 262K | 131K |
| Input $/Mtok | $0.95 | $0.0505 |
| Output $/Mtok | $4.00 | $0.3325 |
| Coding index | 53.4 | Not benchmarked |
| Independent measurements behind it | 4 | 0 |
| Cost per solved task | $0.35 | Not computable |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Kimi K2.7 Code better than Llama 3.2 3B Instruct for coding?
Only Kimi K2.7 Code has a coding score, 53.4 on the Vibe Coding Index; Llama 3.2 3B Instruct has none, so the two cannot be ranked on coding. Pick Llama 3.2 3B Instruct for output tokens 12.0x cheaper ($0.3325 vs $4.00 per million).
Which is cheaper, Kimi K2.7 Code or Llama 3.2 3B Instruct?
Kimi K2.7 Code costs $0.95 per million input tokens and $4.00 per million output tokens; Llama 3.2 3B Instruct costs $0.0505 and $0.3325. Cost per solved task is known for Kimi K2.7 Code only.
Which has the bigger context window, Kimi K2.7 Code or Llama 3.2 3B Instruct?
Kimi K2.7 Code, with 262K tokens against 131K.
How many benchmarks is the Kimi K2.7 Code vs Llama 3.2 3B Instruct comparison based on?
6 benchmarks have a result for both models as of October 21, 2026. Kimi K2.7 Code's coding index rests on 4 independent benchmarks and Llama 3.2 3B Instruct's on 0. Where only one model has a result, the page says so instead of naming a winner.