DeepSeek V3.1 Terminus
DeepSeek V3.1 TerminusAlternatives?
CheaperMiniMax M3.1 Flashthinking×5.5-40%$0.22Complete model response must be appended
BetterGPT-6 Lunalow×2.3-55%$0.005Function calling only with none · Chat Completions+3
OpenMiMo v2.6 Prothinking×5.7+24%$0.13Memory (4-bit): 411 GB → 615 GB+4
CheaperClaude Haiku 5.5max×5.1-55%$0.21Remove sampling parameters+5
CheaperGLM 5.3 FlashmaxOpen×4.6-46%$0.28Defaults to max · Self-hosted+4MiniMax M3.1 Flashthinking×5.5-40%$0.22Complete model response must be appendedClaude Haiku 5.5max×5.1-55%$0.21Remove sampling parameters+5GLM 5.3 FlashmaxOpen×4.6-46%$0.28Defaults to max · Self-hosted+4GPT-6 Lunamax×4.4-55%$0.068Function calling only with none · Chat Completions+3
Qwen3.8 FlashthinkingOpen×4.4-48%$0.37max_tokens range · QwenCloud+5GPT-6 Lunalow×2.3-55%$0.005Function calling only with none · Chat Completions+3GPT-5.6 Lunalow×2.3same$0.01GPT OSS 20BhighOpen+27%-92%$0.011
Nemotron 3 Nano 30B A3BthinkingOpen+32%-76%$0.017MiMo v2.5thinkingOpen×3.1-60%$0.019MiMo v2.6 ProthinkingOpen×5.7+24%$0.13Memory (4-bit): 411 GB → 615 GB+4
Step 5 PreviewthinkingOpen×5.4+224%$1.03web_search not available · Chat Completions+2
Muse Spark OpenxhighOpen×5.1+355%$0.84Memory (4-bit): 411 GB → 493 GB+2DeepSeek V4.1 Flash 1017maxOpen×5+19%$0.26Memory (4-bit): 411 GB → 458 GB+5GLM 5.3maxOpen×4.9+389%$2.01Memory (4-bit): 411 GB → 452 GB+4DeepSeek V4.1 Flash 1017maxOpen×5—$0.26Memory (4-bit): 411 GB → 458 GB+5DeepSeek V3.2 SpecialethinkingOpen×3.3——R1 0528thinkingOpen+83%—$0.39 / $1.63DeepSeek V4 Pro 0813non-thinkingOpen+22%—$0.36Memory (4-bit): 411 GB → 990 GB+5DeepSeek V3.1non-thinkingOpen-23%—$0.25 / $0.95DeepSeek R1 0528 Qwen3 8BthinkingOpen-32%—$0.06 / $0.09Context: 128K → 33KDeepSeek V3 0324non-thinkingOpen-45%—$0.021DeepSeek V3non-thinkingOpen-67%—$0.02All DeepSeek models →Run it yourself
estimateBenchmarks
in Chat & KnowledgeFAQ
How much does DeepSeek V3.1 Terminus cost?›
$0.27 per 1M input tokens and $0.95 per 1M output tokens (cheapest host: DeepInfra).
Is DeepSeek V3.1 Terminus free to use?›
There is no free API; the cheapest paid price is $0.27 per 1M input tokens. The weights are open, so you can run it yourself.
How good is DeepSeek V3.1 Terminus, and what is it best at?›
#208 of 334 in Chat & Knowledge, #75 of 132 in Coding, #93 of 133 in Agents & Tool Use, #64 of 195 in Writing & Language, #81 of 133 in Vision & Documents.
Is DeepSeek V3.1 Terminus better than DeepSeek V3.1?›
On Humanity's Last Exam, DeepSeek V3.1 Terminus scores 16.4 vs 14.3 — higher.
How can I use DeepSeek V3.1 Terminus?›
Through the API of DeepInfra, Featherless, SiliconFlow.
What is the context window of DeepSeek V3.1 Terminus?›
128K tokens, with up to 147K output tokens.
When was DeepSeek V3.1 Terminus released? Is there a newer version?›
Released 2025-09-22. It has been superseded by DeepSeek V3.2.
Is DeepSeek V3.1 Terminus open source?›
Yes, the weights are open (MIT licence).
Can I run DeepSeek V3.1 Terminus locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can DeepSeek V3.1 Terminus read images?›
No, it accepts text only.
What is DeepSeek V3.1 Terminus?›
DeepSeek V3.1 Terminus is an open-weights language model from DeepSeek, released Sep 2025, the successor to DeepSeek V3.1.
What are the best alternatives to DeepSeek V3.1 Terminus?›
MiniMax M3.1 Flash (cheaper), GPT-6 Luna (better), MiMo v2.6 Pro (open weights).