The best open LLMs for your use case:
DeepSeek's frontier 1.6T-parameter Mixture-of-Experts model (49B active per token) with hybrid attention built for long-context, low-cost reasoning. Runs in FP4 with a 512K-token context window.
Speed:
Intelligence:
Price: (1M Tokens)
$1.74 / 3.48Cached input: (1M Tokens)
$0.20Context: (tokens)
512,000Inputs:
Benchmarks:
MRCR 1M
Summarization
SimpleQA
General Knowledge
LiveCodeBench
Coding Agents
SWE-Bench Verified
Coding Agents
Terminal-Bench 2.0
Coding Agents
GDPval-AA
Agents and Function Calling
HLE
General Knowledge
MMLU-Pro
General Knowledge
MRCR 1M
Summarization
SimpleQA
General Knowledge
LiveCodeBench
Coding Agents
SWE-Bench Verified
Coding Agents
Terminal-Bench 2.0
Coding Agents
GDPval-AA
Agents and Function Calling
HLE
General Knowledge
MMLU-Pro
General Knowledge
Try it outNext-generation reasoning model from MiniMax with frontier agentic, coding, and multimodal performance. Strong scores on SWE-Bench, BrowseComp, OmniDocBench, and IMO/USAMO competition reasoning.
Speed:
Intelligence:
Price: (1M Tokens)
$0.30 / 1.20Cached input: (1M Tokens)
$0.06Context: (tokens)
524,288Inputs:
Benchmarks:
GPQA-Diamond
General Knowledge
Video-MME v2
Multimodal - Vision
Claw-Eval
Agents and Function Calling
SWE-Bench Verified
Coding Agents
SWE-Bench Pro
Coding Agents
Apex Agents
Agents and Function Calling
MMMU-Pro
Multimodal - Vision
GPQA-Diamond
General Knowledge
Video-MME v2
Multimodal - Vision
Claw-Eval
Agents and Function Calling
SWE-Bench Verified
Coding Agents
SWE-Bench Pro
Coding Agents
Apex Agents
Agents and Function Calling
MMMU-Pro
Multimodal - Vision
Use case:
Summarization
Features:
Long Context Handling