# Chatgpt 4o Latest 20250326 vs DeepSeek-V2.5 (Sep 2024)

> Chatgpt 4o Latest 20250326 is the stronger model overall, scoring 43.2 to 37.0 on the Noometry Index.

- Canonical page: https://noometry.com/compare/chatgpt-4o-vs-deepseek-v2-5
- Last updated: 2026-10-10
- Shared benchmarks: 17

## Summary

- They share 17 benchmarks with published results for both. Chatgpt 4o Latest 20250326 scores higher in 8 categories and DeepSeek-V2.5 (Sep 2024) in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in writing & preference, where Chatgpt 4o Latest 20250326 leads 62.8 to 50.0.
- DeepSeek-V2.5 (Sep 2024) has downloadable open weights; the other is API-only.

## Snapshot

| | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| Provider | OpenAI | DeepSeek |
| Noometry Index | 43.2 | 37.0 |
| Rank | 84 | 197 |
| Context | — | — |
| Input $/M | — | — |
| Output $/M | — | — |
| Weights | Proprietary | Open |

## Coding

- Chatgpt 4o Latest 20250326: 41.0 (#122)
- DeepSeek-V2.5 (Sep 2024): 31.3 (#274)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Coding | 1413 | 1310 |
| Aider Polyglot | — | 17.8% |
| BigCodeBench Instruct | — | 48.6% |
| BigCodeBench Complete | — | 53.2% |
| HumanEval+ | — | 83.5% |
| MBPP+ | — | 74.1% |

## Reasoning

- Chatgpt 4o Latest 20250326: 32.8 (#64)
- DeepSeek-V2.5 (Sep 2024): 24.5 (#145)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Hard Prompts | 1424 | 1289 |
| Kagi LLM Benchmark | 75% | — |

## Math

- Chatgpt 4o Latest 20250326: 35.8 (#134)
- DeepSeek-V2.5 (Sep 2024): 33.3 (#177)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Math | 1407 | 1289 |

## Knowledge

- Chatgpt 4o Latest 20250326: 38.7 (#131)
- DeepSeek-V2.5 (Sep 2024): 34.2 (#190)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Expert | 1400 | 1266 |
| Confabulations | 16.6% | — |

## Multimodal

- Chatgpt 4o Latest 20250326: 38.9 (#54)
- DeepSeek-V2.5 (Sep 2024): —

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Vision | 1243 | — |

## Multilingual

- Chatgpt 4o Latest 20250326: 53.1 (#72)
- DeepSeek-V2.5 (Sep 2024): 42.7 (#188)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Non-English | 1419 | 1273 |
| LMArena Chinese | 1456 | 1317 |
| LMArena French | 1447 | 1288 |
| LMArena German | 1424 | 1258 |
| LMArena Japanese | 1406 | 1228 |
| LMArena Korean | 1397 | 1209 |
| LMArena Russian | 1429 | 1289 |
| LMArena Spanish | 1435 | 1248 |

## Instruction Following

- Chatgpt 4o Latest 20250326: 74.2 (#102)
- DeepSeek-V2.5 (Sep 2024): 67.7 (#189)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Instruction Following | 1403 | 1280 |

## Long Context

- Chatgpt 4o Latest 20250326: 43.1 (#104)
- DeepSeek-V2.5 (Sep 2024): 39.5 (#169)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Longer Query | 1413 | 1301 |

## Writing & Preference

- Chatgpt 4o Latest 20250326: 62.8 (#68)
- DeepSeek-V2.5 (Sep 2024): 50.0 (#182)

| Benchmark | Chatgpt 4o Latest 20250326 | DeepSeek-V2.5 (Sep 2024) |
|---|---|---|
| LMArena Text | 1429 | 1294 |
| LMArena Creative Writing | 1406 | 1285 |
| LMArena Multi-Turn | 1454 | 1297 |
| EQ-Bench Creative Writing | 1501 | — |

## FAQ

### Is Chatgpt 4o Latest 20250326 better than DeepSeek-V2.5 (Sep 2024)?

Chatgpt 4o Latest 20250326 is the stronger model overall, scoring 43.2 to 37.0 on the Noometry Index.

### Is Chatgpt 4o Latest 20250326 or DeepSeek-V2.5 (Sep 2024) better for coding?

Chatgpt 4o Latest 20250326 scores higher on coding benchmarks: 41.0 versus 31.3 in the Noometry coding category.

### How many benchmarks do Chatgpt 4o Latest 20250326 and DeepSeek-V2.5 (Sep 2024) share?

17 benchmarks have published results for both models. Chatgpt 4o Latest 20250326 has 21 scored results on Noometry and DeepSeek-V2.5 (Sep 2024) has 22.