# Kimi K2.5 vs Llama-3.3-70B-Instruct

> Kimi K2.5 is the stronger model overall, scoring 47.8 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 5.8× less per token, which makes it the better buy when Kimi K2.5's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/kimi-k2-5-vs-llama-3-3-70b-instruct
- Last updated: 2026-10-10
- Shared benchmarks: 26

## Summary

- They share 26 benchmarks with published results for both. Kimi K2.5 scores higher in 9 categories and Llama-3.3-70B-Instruct in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where Kimi K2.5 leads 50.3 to 14.3.
- The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 92.2% for Kimi K2.5 and 5.1% for Llama-3.3-70B-Instruct.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $0.45 / $2.25 for Kimi K2.5.
- Kimi K2.5 accepts more context: 262K tokens versus 128K.

## Snapshot

| | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| Provider | Moonshot AI | Meta |
| Noometry Index | 47.8 | 30.3 |
| Rank | 50 | 283 |
| Context | 262K | 128K |
| Input $/M | $0.45 | $0.10 |
| Output $/M | $2.25 | $0.32 |
| Weights | Open | Open |

## Coding

- Kimi K2.5: 48.7 (#48)
- Llama-3.3-70B-Instruct: 30.6 (#284)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| SciCode | 49% | 26% |
| WeirdML | 45.6% | 14.4% |
| LMArena Coding | 1474 | 1268 |
| SWE-bench Verified | 73.8% | — |
| SWE-bench Verified (bash only) | 70.8% | — |
| LMArena WebDev | 1436 | — |
| SWE-bench Multilingual | 67.3% | — |
| BigCodeBench Instruct | — | 46.9% |
| LiveBench Coding | — | 36.6% |
| BigCodeBench Complete | — | 57.5% |
| ALE-Bench | 821.65 | — |

## Agentic & Tool Use

- Kimi K2.5: 34.0 (#44)
- Llama-3.3-70B-Instruct: 25.6 (#100)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| Terminal-Bench | 43.2% | — |
| Berkeley Function Calling Leaderboard | — | 31.9% |
| OSWorld | 63.3% | — |
| BALROG | — | 23% |
| Vending-Bench 2 | 1,198 | — |

## Reasoning

- Kimi K2.5: 31.2 (#74)
- Llama-3.3-70B-Instruct: 13.7 (#325)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| SimpleBench | 46.8% | 19.9% |
| CritPt | 3.1% | 0% |
| LMArena Hard Prompts | 1454 | 1257 |
| Epoch Capabilities Index | 148.05 | 127.32 |
| ARC-AGI-2 | 11.8% | — |
| Kagi LLM Benchmark | 78.5% | — |
| NYT Connections (extended) | 69.9% | — |
| ARC-AGI-1 | 65.3% | — |
| Chess Puzzles | 12% | — |
| EnigmaEval | 3.4% | — |
| Thematic Generalization | 69.4% | — |
| LiveBench Reasoning | — | 50.8% |
| DTBench | — | 59.5% |
| LiveBench Data Analysis | — | 49.5% |
| LMCA | — | 17.5% |
| ForecastBench | — | 58.6 |
| LiveBench | — | 50.2% |

## Math

- Kimi K2.5: 50.3 (#46)
- Llama-3.3-70B-Instruct: 14.3 (#292)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 92.2% | 5.1% |
| LMArena Math | 1470 | 1268 |
| MathArena Final-Answer Competitions | 62.3% | — |
| LiveBench Math | — | 42.2% |
| MATH Level 5 | — | 41.6% |
| FrontierMath (Feb 2025 set) | 27.9% | — |
| FrontierMath Tier 4 (v1) | 4.2% | — |

## Knowledge

- Kimi K2.5: 53.6 (#52)
- Llama-3.3-70B-Instruct: 30.5 (#220)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | 87.6% | 47.4% |
| Vectara Hallucination Rate | 14.2% | 4.1% |
| LMArena Expert | 1466 | 1225 |
| Humanity's Last Exam | 24.4% | — |
| SimpleQA Verified | 34.3% | — |
| Confabulations | — | 22.8% |
| MMLU | — | 86.3% |

## Multimodal

- Kimi K2.5: 40.4 (#37)
- Llama-3.3-70B-Instruct: —

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Vision | 1269 | — |
| LMArena Document | 1430 | — |

## Multilingual

- Kimi K2.5: 54.1 (#50)
- Llama-3.3-70B-Instruct: 40.1 (#215)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | 1433 | 1236 |
| LMArena Chinese | 1495 | 1217 |
| LMArena French | 1454 | 1280 |
| LMArena German | 1441 | 1251 |
| LMArena Japanese | 1422 | 1149 |
| LMArena Korean | 1412 | 1142 |
| LMArena Russian | 1436 | 1252 |
| LMArena Spanish | 1450 | 1269 |

## Instruction Following

- Kimi K2.5: 75.5 (#60)
- Llama-3.3-70B-Instruct: 71.3 (#152)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Instruction Following | 1431 | 1242 |
| LiveBench Instruction Following | — | 82.7% |

## Long Context

- Kimi K2.5: 52.1 (#7)
- Llama-3.3-70B-Instruct: 26.4 (#290)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| Fiction.LiveBench | 86.1% | 33.3% |
| LMArena Longer Query | 1445 | 1256 |
| CL-bench | 19.3% | — |
| CL-bench Life | 13.2% | — |

## Writing & Preference

- Kimi K2.5: 65.3 (#51)
- Llama-3.3-70B-Instruct: 47.8 (#203)

| Benchmark | Kimi K2.5 | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | 1445 | 1274 |
| LMArena Creative Writing | 1423 | 1250 |
| LMArena Multi-Turn | 1444 | 1279 |
| EQ-Bench Creative Writing | 1579 | — |
| LiveBench Language | — | 39.2% |

## FAQ

### Is Kimi K2.5 better than Llama-3.3-70B-Instruct?

Kimi K2.5 is the stronger model overall, scoring 47.8 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 5.8× less per token, which makes it the better buy when Kimi K2.5's lead doesn't matter for your workload.

### Which is cheaper, Kimi K2.5 or Llama-3.3-70B-Instruct?

Llama-3.3-70B-Instruct is cheaper. It lists at $0.10 per million input tokens and $0.32 per million output tokens; Kimi K2.5 lists at $0.45 and $2.25.

### Is Kimi K2.5 or Llama-3.3-70B-Instruct better for coding?

Kimi K2.5 scores higher on coding benchmarks: 48.7 versus 30.6 in the Noometry coding category.

### Which has the bigger context window?

Kimi K2.5 does, with 262K tokens against 128K.

### How many benchmarks do Kimi K2.5 and Llama-3.3-70B-Instruct share?

26 benchmarks have published results for both models. Kimi K2.5 has 51 scored results on Noometry and Llama-3.3-70B-Instruct has 43.