Skip to content

Compare models

Put up to 4 models beside each other — token prices, context windows, capabilities and provider, from the same catalogue the model pages read.

  1. Gemini 4 CarbonGoogleRemove
  2. DeepSeek V4.1 Flash 1017DeepSeekRemove
  3. GPT OSS 20BOpenAIRemove
gemini-4-carbon vs deepseek-v4.1-flash-1017 vs gpt-oss-20b
AttributeGemini 4 Carbongemini-4-carbonDeepSeek V4.1 Flash 1017deepseek-v4.1-flash-1017GPT OSS 20Bgpt-oss-20b
Pricing
Input$2.00 / 1M$0.20 / 1M$0.075 / 1M
Output$10.00 / 1M$0.80 / 1M$3.00 / 1M
Cache Write (5m)$2.00 / 1M$0.20 / 1M$0.075 / 1M
Cache Write (1h)$2.00 / 1M$0.20 / 1M$0.075 / 1M
Cache Read$2.00 / 1M$0.20 / 1M$0.075 / 1M
Web Search$0 / 1M$0 / 1M$0 / 1M
Context
Max context1M1M131.1K
Max outputN/AN/AN/A
Capabilities
VisionYesYesNo
Function CallingYesYesYes
JSON ModeYesYesYes
StreamingYesYesYes
Catalogue
ProviderGoogleDeepSeekOpenAI
Categorychatchatchat
Charge typePay As You GoPay As You GoPay As You Go
Released———
Description
SummaryGemini 4 Carbon is Google's frontier multimodal reasoning model, designed for advanced coding, agentic workflows, and complex multi-step problem solving. It supports text, image, audio, video, and file inputs, enabling integrated reasoning and analysis across diverse modalities. Optimized for long-horizon execution and demanding professional workloads, Gemini 4 Carbon supports up to 1M output tokens through continuation and configurable reasoning effort levels (low, medium, high) to balance reasoning depth, latency, and cost. It is well suited for autonomous agents, software engineering, multimodal analysis, and large-scale workflow automation.DeepSeek V4.1 Flash 1017 is the October 17, 2026 snapshot of DeepSeek’s cost-efficient Mixture-of-Experts (MoE) model, built on its innovative Causal Encoder-Decoder (CED) architecture. Featuring 552B total parameters with 8B activated for input and 16B for output, it combines efficient inference with native multimodal understanding through jointly trained text and visual representations. Optimized for coding, terminal and computer-use agents, and long-horizon workflows, the model introduces compressed KV caching that reduces cache memory usage to approximately one-quarter of the previous Flash generation. DeepSeek reports that it surpasses V4 Pro in performance, inference speed, and task completion time, making it well suited for cost-efficient autonomous agents, complex software engineering, and long-context analysis.gpt-oss-20b is an open-weight, 21B-parameter OpenAI model released under Apache 2.0. It uses a Mixture-of-Experts design so only ~3.6B parameters run each step, enabling faster, lower-cost inference on consumer or single-GPU hardware. Trained in the Harmony format, it supports configurable reasoning depth, fine-tuning, function calling, tool use, and structured outputs.