# GPT-5.2 Codex vs Llama-3.3-70B-Instruct

> GPT-5.2 Codex is the stronger model overall, scoring 42.2 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 31× less per token, which makes it the better buy when GPT-5.2 Codex's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/gpt-5-2-codex-vs-llama-3-3-70b-instruct
- Last updated: 2026-10-10
- Shared benchmarks: 0

## Summary

- The widest gap is in agentic & tool use, where GPT-5.2 Codex leads 40.7 to 25.6.
- Llama-3.3-70B-Instruct is cheaper at $0.10 / $0.32 per million input/output tokens, against $1.75 / $14 for GPT-5.2 Codex.
- GPT-5.2 Codex accepts more context: 400K tokens versus 128K.
- Llama-3.3-70B-Instruct has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| Provider | OpenAI | Meta |
| Noometry Index | 42.2 | 30.3 |
| Rank | 107 | 283 |
| Context | 400K | 128K |
| Input $/M | $1.75 | $0.10 |
| Output $/M | $14 | $0.32 |
| Weights | Proprietary | Open |

## Coding

- GPT-5.2 Codex: 45.4 (#66)
- Llama-3.3-70B-Instruct: 30.6 (#284)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| SWE-bench Verified (bash only) | 72.8% | — |
| LMArena WebDev | 1339 | — |
| SWE-bench Multilingual | 66.3% | — |
| SciCode | — | 26% |
| WeirdML | — | 14.4% |
| BigCodeBench Instruct | — | 46.9% |
| LiveBench Coding | — | 36.6% |
| LMArena Coding | — | 1268 |
| BigCodeBench Complete | — | 57.5% |
| ALE-Bench | 1,300 | — |

## Agentic & Tool Use

- GPT-5.2 Codex: 40.7 (#20)
- Llama-3.3-70B-Instruct: 25.6 (#100)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| Terminal-Bench | 66.5% | — |
| Berkeley Function Calling Leaderboard | — | 31.9% |
| BALROG | — | 23% |

## Reasoning

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 13.7 (#325)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| SimpleBench | — | 19.9% |
| CritPt | — | 0% |
| LiveBench Reasoning | — | 50.8% |
| LMArena Hard Prompts | — | 1257 |
| DTBench | — | 59.5% |
| LiveBench Data Analysis | — | 49.5% |
| LMCA | — | 17.5% |
| Epoch Capabilities Index | — | 127.32 |
| ForecastBench | — | 58.6 |
| LiveBench | — | 50.2% |

## Math

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 14.3 (#292)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| OTIS Mock AIME 2024-2025 | — | 5.1% |
| LiveBench Math | — | 42.2% |
| LMArena Math | — | 1268 |
| MATH Level 5 | — | 41.6% |

## Knowledge

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 30.5 (#220)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| GPQA Diamond | — | 47.4% |
| Confabulations | — | 22.8% |
| Vectara Hallucination Rate | — | 4.1% |
| LMArena Expert | — | 1225 |
| MMLU | — | 86.3% |

## Multilingual

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 40.1 (#215)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Non-English | — | 1236 |
| LMArena Chinese | — | 1217 |
| LMArena French | — | 1280 |
| LMArena German | — | 1251 |
| LMArena Japanese | — | 1149 |
| LMArena Korean | — | 1142 |
| LMArena Russian | — | 1252 |
| LMArena Spanish | — | 1269 |

## Instruction Following

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 71.3 (#152)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| LiveBench Instruction Following | — | 82.7% |
| LMArena Instruction Following | — | 1242 |

## Long Context

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 26.4 (#290)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| Fiction.LiveBench | — | 33.3% |
| LMArena Longer Query | — | 1256 |

## Writing & Preference

- GPT-5.2 Codex: —
- Llama-3.3-70B-Instruct: 47.8 (#203)

| Benchmark | GPT-5.2 Codex | Llama-3.3-70B-Instruct |
|---|---|---|
| LMArena Text | — | 1274 |
| LMArena Creative Writing | — | 1250 |
| LMArena Multi-Turn | — | 1279 |
| LiveBench Language | — | 39.2% |

## FAQ

### Is GPT-5.2 Codex better than Llama-3.3-70B-Instruct?

GPT-5.2 Codex is the stronger model overall, scoring 42.2 to 30.3 on the Noometry Index. Llama-3.3-70B-Instruct costs 31× less per token, which makes it the better buy when GPT-5.2 Codex's lead doesn't matter for your workload.

### Which is cheaper, GPT-5.2 Codex 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; GPT-5.2 Codex lists at $1.75 and $14.

### Is GPT-5.2 Codex or Llama-3.3-70B-Instruct better for coding?

GPT-5.2 Codex scores higher on coding benchmarks: 45.4 versus 30.6 in the Noometry coding category.

### Which has the bigger context window?

GPT-5.2 Codex does, with 400K tokens against 128K.

### How many benchmarks do GPT-5.2 Codex and Llama-3.3-70B-Instruct share?

0 benchmarks have published results for both models. GPT-5.2 Codex has 5 scored results on Noometry and Llama-3.3-70B-Instruct has 43.