Skip to content
Quantoros

Compare the major reasoning models: o1, o3, DeepSeek-R1, Claude 3.7, Gemini 2.0 Flash Thinking.

LLM & AILevel: IntermediateKey Models & Comparisons

Compare the major reasoning models: o1, o3, DeepSeek-R1, Claude 3.7, Gemini 2.0 Flash Thinking.

Hints
  • Identify the stage: pretraining, adaptation, generation or evaluation.
  • Name the inputs, constraints and a case where the approach fails.

Answer

ModelProviderOpen SourceThinking VisibleKey Strength
o1OpenAINoNo (hidden)First major reasoning model; strong math/code
o3OpenAINoNoState-of-art on benchmarks; high cost
o4-miniOpenAINoNoEfficient reasoning; good cost/perf ratio
DeepSeek-R1DeepSeekYesYesOpen weights; matches o1 at fraction of cost
Claude 3.7 SonnetAnthropicNoYes (optional)Extended thinking; strong coding
Gemini 2.0 Flash ThinkingGoogleNoYesFast; multimodal reasoning
Qwen QwQAlibabaYesYesStrong math; open weights

DeepSeek-R1's significance: it proved RLVR reasoning training works with open-source models, democratizing the technique. R1's distilled variants (1.5B–70B) brought reasoning capabilities to consumer hardware.

Source and licence: JustInCache & contributors — Awesome AI Interviews · Apache-2.0

Related questions