Home / Models / Kryyt AI

DeepSeek · MIT

Kryyt AI API

Frontier open reasoning and agentic coding at commodity prices.

Get API keyCompare models

Context: 1M tokens · Released: September 2026

# Kryyt AI via kryyt (OpenAI SDK)
from openai import OpenAI
client = OpenAI(base_url="https://api.kryyt.com/v1", api_key="$KRYYT_API_KEY")
r = client.chat.completions.create(
    model="deepseek-v4.1-flash",
    messages=[{"role":"user","content":"Explain MoE routing in one paragraph."}],
)
print(r.choices[0].message.content)

Overview

Kryyt AI currently uses DeepSeek V4.1 Flash, a 552B-parameter mixture-of-experts model with native image understanding, a 1M-token context, up to 384K output tokens, and a controllable reasoning-effort setting. DeepSeek reports it outperforms the previous V4 Pro on its own tests. V4.1 Pro is the upcoming higher-capability tier.

Best for:

  • Long-context code generation
  • Step-by-step reasoning
  • High-volume, cost-sensitive workloads
Kryyt AIExample Kryyt AI workspace with chat, project context, and linked files

DeepSeek V4.1 benchmark performance

DeepSeek V4.1 Flash — live on kryyt today — posts the highest vendor-reported Terminal-Bench 2.1 and DeepSWE scores of any open model in our roster, and a 3,471 Codeforces rating, up from 3,289 for V4 Flash.

90.9GPQA Diamond
90.6Terminal-Bench 2.1
74.2DeepSWE v1.1
3,471Codeforces rating
DeepSeek V4.1 evaluation results
BenchmarkModelScore (%)
GPQA DiamondGraduate-level science reasoningDeepSeek V4.1 Flash
90.9
Terminal-Bench 2.1Agentic coding in real terminal environmentsDeepSeek V4.1 Flash
90.6
CyberGymSecurity vulnerability discoveryDeepSeek V4.1 Flash
88.1
DeepSWE v1.1Repairing fresh, real-world software issuesDeepSeek V4.1 Flash
74.2
NL2Repo-BenchGenerating whole repositories from a specDeepSeek V4.1 Flash
65.4
AutomationBenchWorkflow-automation agentsDeepSeek V4.1 Flash
54.8
Agents' Last ExamLong-horizon agentic tasksDeepSeek V4.1 Flash
31.8

Sources: DeepSeek-V4.1-Flash model card (NVIDIA) (opens in new tab); DeepSeek V4.1 Flash review (Build Fast with AI) (opens in new tab).

DeepSeek V4.1 Flash vs DeepSeek V4 Flash

Generation-over-generation: DeepSeek V4 Flash → DeepSeek V4.1 Flash
BenchmarkScores (%)
DeepSWE v1.1Repairing fresh, real-world software issues
DeepSeek V4 Flash54.4
DeepSeek V4.1 Flash74.2
Terminal-Bench 2.1Agentic coding in real terminal environments
DeepSeek V4 Flash82.7
DeepSeek V4.1 Flash90.6
GPQA DiamondGraduate-level science reasoning
DeepSeek V4 Flash89.9
DeepSeek V4.1 Flash90.9

Sources: DeepSeek-V4.1-Flash technical report (opens in new tab); DeepSeek V4.1 Flash review (Build Fast with AI) (opens in new tab).

DeepSeek on GPQA Diamond, generation by generation

DeepSeek on GPQA Diamond, generation by generationGPQA Diamond (%) — V3 (Dec 2024): 59.1; R1 (Jan 2025): 71.5; R1-0528 (May 2025): 81.0; V4 Flash (Apr 2026): 89.9; V4.1 Flash (Sep 2026): 90.95060708090100GPQA Diamond (%)59.1V3Dec 202471.5R1Jan 202581.0R1-0528May 202589.9V4 FlashApr 202690.9V4.1 FlashSep 2026
DeepSeek on GPQA Diamond, generation by generation. Sources: DeepSeek-V3 technical report (opens in new tab); DeepSeek API change log (opens in new tab); DeepSeek-V4.1-Flash technical report (opens in new tab).
View chart data as a table
ModelReleasedGPQA Diamond (%)
V3Dec 202459.1
R1Jan 202571.5
R1-0528May 202581.0
V4 FlashApr 202689.9
V4.1 FlashSep 202690.9

Scores are vendor-reported unless marked independent, as of September 2026. Labs use different agent harnesses, reasoning settings, and benchmark versions, so treat cross-lab gaps of a few points as roughly equal and test on your own workload. See all models on the benchmarks page.

Kryyt AI models & pricing

ModelAPI IDContextPricingLicenseStatus
Kryyt AIdeepseek-v4.1-flash1MSee account pricingMITLive
DeepSeek V4.1 Prodeepseek-v4.1-pro1MSee account pricingMITSoon

deepseek-v4.1-flash

Model
Kryyt AI
Context
1M
Pricing
See account pricing
License
MIT
Status
Live

deepseek-v4.1-pro

Model
DeepSeek V4.1 Pro
Context
1M
Pricing
See account pricing
License
MIT
Status
Soon

Usage is charged at Kryyt's published rates for the selected model and features. Review current pricing and usage in your account.

Kryyt AI FAQ

What is DeepSeek V4.1?

DeepSeek V4.1 Flash is a 552B-parameter mixture-of-experts model with native image understanding, a 1M-token context, up to 384K output tokens, and a controllable reasoning-effort setting. DeepSeek reports it outperforms the previous V4 Pro on its own tests. V4.1 Pro is the upcoming higher-capability tier.

How do I call Kryyt AI on kryyt?

Send a POST request to https://api.kryyt.com/v1/chat/completions with your kryyt key and set "model" to one of: deepseek-v4.1-flash, deepseek-v4.1-pro.

What license is DeepSeek V4.1 under?

DeepSeek V4.1 is available under the MIT license via its provider, DeepSeek.

What context window does DeepSeek V4.1 support?

DeepSeek V4.1 supports 1M tokens depending on the specific model variant.

Related models

Z.ai (Zhipu)

GLM 5.3

Agentic engineering — long-horizon coding, terminal tasks, hybrid reasoning.

Moonshot AI

Kimi K3

The largest open-weights model — native vision and multi-agent Swarm.

Alibaba Cloud

Qwen3.8

Alibaba's most capable family — multimodal flagship plus an open dense model.

Ship with one endpoint

Get a key, change your base URL, and call every model on this page. No new SDKs.

Get your API key