DeepSeek V3.2
Language AvailableCompare models
ComparingDeepSeek V3.2with
Pick a model above to see the comparison.
deepseek/deepseek-v3.2 · by DeepSeek · mixture-of-experts
Pricing — 1 offering(s)
Input tokens
- $0.28 / 1M tokens (input) Current 2025-12-01 → present
Output tokens
- $0.42 / 1M tokens (output) Current 2025-12-01 → present
Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.
Capability profile
reasoning strong
coding moderate
tool use strong
instruction following strong
context window moderate
multilingual moderate
speed moderate
cost efficiency strong
Coding benchmarks
View full leaderboard →| Benchmark | Score | Harness | Source |
|---|---|---|---|
| swe-bench-verified ↗ | 73.1% 2025-12 | internal | 📄 paper ↗ |
Operator guidance
Exceptional price-to-performance ratio, same niche V3 held. Route here when cost is the primary constraint and reasoning/tool-use capability is still required. Note first-party availability is in transition: DeepSeek's live pricing page moved to V4-preview pricing on 2026-04-24 and the legacy deepseek-chat/deepseek-reasoner aliases deprecate 2026-07-24; open weights remain available for self-hosting and third-party hosts.
Use cases
- Cost-optimised general reasoning and code generation
- Agentic workflows needing tool calls inside thinking mode
- Open-weights deployment on own infrastructure
Limitations
- Chinese lab; some organisations have data-residency or compliance constraints
- Open weights means self-hosting cost must be factored for on-prem use
- Capability ratings are qualitative, not a single cited benchmark run
- First-party serving beyond the 2026-07-24 V4 alias cutover is unconfirmed