# DeepSeek V3.2 Exp

DeepSeek V3.2 Exp costs $0.27 per million input tokens and $0.41 per million output tokens, as of October 22, 2026. Only one of the 6 coding benchmarks behind the index has measured it, and the index needs two, so there is not enough data for a coding score. 58 benchmarks outside the index have a result below.

Page: https://llmbenchmark.ai/models/deepseek-v3-2-exp
Data from October 22, 2026. Every figure below names the board it was read from and the day.

- Maker: DeepSeek
- Released: September 29, 2025
- Price per million input tokens: $0.27
- Price per million output tokens: $0.41
- Context window: 164K tokens
- Vibe Coding Index: not enough data (one index benchmark; at least two are needed to rank)

Benchmark results (59 of 195):

- Aider polyglot: 74.2% (source: Aider polyglot leaderboard, read October 22, 2026)
- LMArena WebDev: 1272 Elo (source: LMArena WebDev, read October 22, 2026)
- AA Intelligence Index: 16.6 (source: Artificial Analysis (via OpenRouter), read October 22, 2026)
- Design Arena: websites: 1178 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: UI components: 1164 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: game dev: 1144 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: data viz: 1152 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: 3D: 1148 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: all code categories: 1170 Elo (source: Design Arena, read October 22, 2026)
- Design Arena: SVG: 1031 Elo (source: Design Arena, read October 22, 2026)
- AA Math Index: 87.7 (source: Artificial Analysis, read October 22, 2026)
- GPQA Diamond (Artificial Analysis run): 79.7% (source: Artificial Analysis, read October 22, 2026)
- Humanity's Last Exam (Artificial Analysis run): 14.9% (source: Artificial Analysis, read October 22, 2026)
- MMLU Pro (Artificial Analysis run): 85.0% (source: Artificial Analysis, read October 22, 2026)
- LiveCodeBench (Artificial Analysis run): 78.9% (source: Artificial Analysis, read October 22, 2026)
- Terminal-Bench Hard (Artificial Analysis run): 31.1% (source: Artificial Analysis, read October 22, 2026)
- IFBench: 54.1% (source: Artificial Analysis, read October 22, 2026)
- Long context reasoning (AA-LCR): 72.3% (source: Artificial Analysis, read October 22, 2026)
- τ²-Bench Telecom: 33.9% (source: Artificial Analysis, read October 22, 2026)
- AIME 2025 (Artificial Analysis run): 87.7% (source: Artificial Analysis, read October 22, 2026)
- WeirdML: 39.5% (source: Epoch AI benchmarking hub, read October 22, 2026)
- DTBench: 87.7% (source: Epoch AI benchmarking hub, read October 22, 2026)
- LMCA: 29.1% (source: Epoch AI benchmarking hub, read October 22, 2026)
- Fiction.LiveBench: 83.3% (source: Epoch AI benchmarking hub, read October 22, 2026)
- CL-bench: 13.2% (source: Epoch AI benchmarking hub, read October 22, 2026)
- TheAgentCompany: 42.9% (source: Epoch AI benchmarking hub, read October 22, 2026)
- CL-bench Life: 9.5% (source: Epoch AI benchmarking hub, read October 22, 2026)
- Epoch Capabilities Index: 144.9 (source: Epoch AI benchmarking hub, read October 22, 2026)
- LMArena Text: 1425 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: coding: 1438 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: hard prompts: 1429 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: math: 1422 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: instruction following: 1401 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: creative writing: 1404 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: English: 1441 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: non-English: 1404 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Chinese: 1461 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: German: 1440 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Japanese: 1331 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Korean: 1372 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: French: 1434 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Spanish: 1441 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Russian: 1412 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: Polish: 1435 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: multi-turn: 1422 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: longer queries: 1418 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: expert: 1422 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: hard prompts in English: 1444 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: software and IT: 1438 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: mathematical work: 1433 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: science: 1443 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: medicine and healthcare: 1468 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: legal and government: 1419 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: business and finance: 1412 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: writing and language: 1407 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena Text: entertainment and media: 1407 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- LMArena WebDev: HTML: 1281 Elo (source: LMArena (Hugging Face dataset), read October 22, 2026)
- Berkeley Function Calling Leaderboard (BFCL): 56.7% (source: Berkeley Function Calling Leaderboard, read October 22, 2026)
- Kagi LLM Benchmark: 52.2% (source: Kagi LLM Benchmarking Project, read October 22, 2026)