DeepSeek V3.2
DeepSeekOpen weightsMITReleased Dec 1, 2025
DeepSeek V3.2 is not on the arena.ai Agents leaderboard, so it has no rank on our default leaderboard; the other data below is for reference. DeepSeek V3.2 ranks #127 on the LMArena Text leaderboard (style control) with an Arena score of 1424.6 (±3.5, 48,095 votes). API price: $0.28 input / $0.42 output per 1M tokens. Context: 164K. Nearest rivals: Claude Haiku 5.5 (1424.8) above and GLM-4.6 (1424.3) below.
Quick answer
- Agents leaderboard
- No result yet
- Text (LMArena)
- #127 · 1424.6
- API price (in / out)
- $0.28 / $0.42 per 1M tokens
- Closest rival
- Claude Haiku 5.5 · 1424.8
- Cheaper alternative
- DeepSeek V4 Flash: Text 1435.7, output $0.28 (33% cheaper)
- Data updated
- Oct 29, 2026 · synced daily
Sources: arena.ai, LMArena, Artificial Analysis, OpenRouter
Key metrics
Rank history
LMArena rank, last 14 days#123 on Oct 16, #127 now; best #123 in this period.
Models ranked nearby
Sorted by Arena score| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 58 |  MiMo V2.5 | 1433.9 | – | $0.27 | – |
| 59 |  Mistral Medium 3.5 | 1426.4 | – | $7.50 | 157 |
| 60 |  Claude Haiku 5.5 | 1424.8 | 36.2 | $0.50 | 236 |
| 61 | DeepSeek V3.2This page | 1424.6 | – | $0.42 | – |
| 62 |  GLM-4.6 | 1424.3 | – | $2.20 | – |
| 63 |  Qwen3 Max | 1423.5 | – | – | – |
More from DeepSeek
| # | Model | Arena | Intel. | Output $ | tok/s |
|---|---|---|---|---|---|
| 29 | DeepSeek V4.1 Flash | 1476.2 | – | $1.20 | – |
| 42 | DeepSeek V4 Pro | 1465.1 | 23.1 | $3.96 | 76 |
| 57 | DeepSeek V4 Flash | 1435.7 | – | $0.28 | – |
Pricing
- Official API list price: $0.28 input and $0.42 output per 1M tokens (USD).
- At a typical 3:1 input-to-output ratio, the blended cost is about $0.32 per 1M tokens.
- Among 67 models with an output price, it is the #3 cheapest.
- The cheapest OpenRouter route charges $0.28 input / $0.42 output.
Sources & methodology
DeepSeek V3.2 is not on the arena.ai Agents leaderboard, so it has no rank on our default leaderboard; the other data below is for reference.
- arena.ai Agents — agent Net Improvement, 95% interval, sessions and rank (via the public Hugging Face dataset, CC BY 4.0).
- LMArena — Arena score (Text leaderboard with style control, the same view arena.ai shows by default), 95% confidence interval, votes and rank.
- Artificial Analysis — Intelligence / Coding Index, output speed and latency (medians).
- OpenRouter — Route prices, context length and listing date.
Arena scores use the Bradley-Terry model (not Elo); ± is the half-width of the 95% confidence interval. Latency figures are Artificial Analysis medians; for reasoning models, time to first token and to first answer include the thinking phase. Data syncs daily at 14:20 (UTC+8).
Limits: Arena scores come from anonymous users voting on chats, so they reflect conversational preference rather than agent or coding benchmarks; with few votes the interval is wide and ranks move. Prices and speeds change as providers update them.
FAQ
What rank is DeepSeek V3.2 right now?
DeepSeek V3.2 is #61 on our leaderboard of 63 scored models (Arena score 1424.6); its raw LMArena rank is #127.
What does DeepSeek V3.2 cost via API?
Input $0.28 / output $0.42 per 1M tokens (official API list price; the OpenRouter quote when no official price is available). At a typical 3:1 input-to-output ratio, the blended cost is about $0.32 per 1M tokens.
How close is DeepSeek V3.2 to the models ranked next to it?
It is 0.2 points behind Claude Haiku 5.5 (1424.8) and 0.3 points ahead of GLM-4.6 (1424.3); gaps smaller than the confidence interval (±3.5) are not significant.
Is DeepSeek V3.2 open source? Can it be self-hosted?
DeepSeek V3.2 is an open-weights model (license: MIT); you can download the weights and self-host.
What is the context length of DeepSeek V3.2?
164K (163,840 tokens, as listed on OpenRouter).
DeepSeek V3.2 vs Claude Haiku 5.5: which is better?
On the LMArena Text leaderboard DeepSeek V3.2 scores 1424.6 and Claude Haiku 5.5 scores 1424.8 (a 0.2-point gap, within the confidence interval, so not significant). On price, DeepSeek V3.2 costs $0.42 and Claude Haiku 5.5 $0.50 per 1M output tokens.
Is there a cheaper model close to DeepSeek V3.2?
Yes: DeepSeek V4 Flash scores 1435.7 on the Text leaderboard (DeepSeek V3.2: 1424.6) and costs $0.28 per 1M output tokens, about 33% less than DeepSeek V3.2 ($0.42).
Where does the DeepSeek V3.2 data come from and how often is it updated?
Agent success rates come from the arena.ai Agents leaderboard (CC BY 4.0), Text scores from LMArena, intelligence and speed from Artificial Analysis, and prices and context from official list prices and OpenRouter. 17NAS syncs daily at 06:20 UTC; this page was last updated Oct 29, 2026.