ZMIME
Comparison · 3 models · Updated Nov 18, 2026

DeepSeek-R1 vs GPT-4.1 vs Claude Opus 4

Too close to call on our weighted score (Claude Opus 4 66, GPT-4.1 65, DeepSeek-R1 56). The right pick depends on what you value most.

Price in rating: Excluded · Include price

  1. DeepSeek

    DeepSeek-R1

    Released Jan 20, 2025

    56/100
    • ECI138.6
    • Price$1.02 / $5.14
    • Context128K
  2. OpenAI

    GPT-4.1

    Released Apr 14, 2025

    65/100
    • ECI136.2
    • Price$2.00 / $8.00
    • Context1.05M
  3. Anthropic

    Claude Opus 4

    Released May 22, 2025

    66/100
    • ECI142.9
    • Price$15.00 / $75.00
    • Context200K
01 — Verdict

Too close to call

It is close. Our weighted score puts them within a point (Claude Opus 4 66/100, GPT-4.1 65/100, DeepSeek-R1 56/100), so choose by what matters most for your work: Claude Opus 4 for raw capability, DeepSeek-R1 on price and GPT-4.1 for long inputs. The score weighs capability 67%, inputs & features 20%, context window 13%.

  • CapabilityClaude Opus 4Capabilities Index (ECI): Claude Opus 4 142.9 · DeepSeek-R1 138.6 · GPT-4.1 136.2
  • Lowest priceDeepSeek-R1DeepSeek-R1 $2.05 · GPT-4.1 $3.50 · Claude Opus 4 $30.00 per 1M tokens (3:1 blend)
  • Longest contextGPT-4.1GPT-4.1 1,047,576 · Claude Opus 4 200,000 · DeepSeek-R1 128,000 tokens
  • Widest inputsGPT-4.1 and Claude Opus 4DeepSeek-R1: Text · GPT-4.1: Text, Images, PDFs · Claude Opus 4: Text, Images, PDFs
  • Self-hostingDeepSeek-R1Publishes downloadable weights
How the score is built
MeasureWeightDeepSeek-R1GPT-4.1Claude Opus 4
CapabilityCapabilities Index (ECI)67%676471
Inputs & features20%357070
Context window13%326539
Overall100%56/10065/10066/100
02 — Side by side

Every spec in one table

Highlighted cells lead their row. Dashes mean the data is not published.

DeepSeek-R1 vs GPT-4.1 vs Claude Opus 4 specifications side by side
SpecificationDeepSeek-R1DeepSeekGPT-4.1OpenAIClaude Opus 4Anthropic
Capability
Capabilities Index (ECI)138.6136.2142.9 (best)
ECI rank#118 of 158#123 of 158#101 of 158 (best)
GPQA DiamondGraduate-level science questions71.7%66.9%76.3% (best)
FrontierMath Tiers 1–3Research-level mathematics—6.0%—
OTIS Mock AIME 2024–2025Competition mathematics53.3%38.3%64.4% (best)
SWE-bench VerifiedFixing real GitHub issues—48.5%70.7% (best)
SimpleQA VerifiedShort factual questions—31.1%—
Price per million tokens
Input$1.02 (best)$2.00$15.00
Output$5.14 (best)$8.00$75.00
Cached input—$0.50—
Blended (3:1)$2.05 (best)$3.50$30.00
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 12 providersOfficial OpenAI APIMedian of 7 providers
Limits
Context window128,000 tokens1,047,576 tokens (best)200,000 tokens
Max output32,768 tokens (best)32,768 tokens (best)32,000 tokens
Inputs and features
TextYesYesYes
ImagesNoYesYes
PDFsNoYesYes
AudioNoNoNo
VideoNoNoNo
ReasoningYesNoYes
Tool callingYesYesYes
Structured outputNoYesNo
Availability
WeightsOpenProprietaryProprietary
API model ID—gpt-4.1—
API providers1332 (best)8
ReleasedJan 20, 2025Apr 14, 2025May 22, 2025
Knowledge cutoffJul 2024Apr 2024Mar 31, 2025
03 — Cost

What would a month cost?

Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.

  • DeepSeek-R1$20.53
  • GPT-4.1$36.00
  • Claude Opus 4$300.00
04 — Questions

Which should you choose?

Which is better: DeepSeek-R1, GPT-4.1 or Claude Opus 4?

It is close. Our weighted score puts them within a point (Claude Opus 4 66/100, GPT-4.1 65/100, DeepSeek-R1 56/100), so choose by what matters most for your work: Claude Opus 4 for raw capability, DeepSeek-R1 on price and GPT-4.1 for long inputs. The score weighs capability 67%, inputs & features 20%, context window 13%.

Which is cheaper, DeepSeek-R1, GPT-4.1 or Claude Opus 4?

DeepSeek-R1 is cheaper at $1.02 input / $5.14 output per million tokens (median across 12 API providers). GPT-4.1 costs $2.00 input / $8.00 output per million tokens (official OpenAI API price); Claude Opus 4 costs $15.00 input / $75.00 output per million tokens (median across 7 API providers). At a typical mix of three input tokens to one output token, that is $2.05 per million tokens for DeepSeek-R1 versus $3.50 for GPT-4.1 (1.7× as much) and $30.00 for Claude Opus 4 (15× as much).

Which scores higher on benchmarks?

Claude Opus 4 scores higher on the Capabilities Index (ECI): Claude Opus 4 142.9 (#101 of 158), DeepSeek-R1 138.6 (#118 of 158) and GPT-4.1 136.2 (#123 of 158). Their confidence ranges do not overlap (141.3–144.5 vs 136.7–140.4), so the gap is a real one. On individual benchmarks: GPQA Diamond — Claude Opus 4 76.3%, DeepSeek-R1 71.7%, GPT-4.1 66.9%; OTIS Mock AIME 2024–2025 — Claude Opus 4 64.4%, DeepSeek-R1 53.3%, GPT-4.1 38.3%.

Which is better for coding?

There are no published SWE-bench Verified results for DeepSeek-R1 yet, so there is no like-for-like coding score. On overall capability, Claude Opus 4 leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.

Which has the bigger context window?

GPT-4.1 has the largest context window at 1,047,576 tokens, against 200,000 for Claude Opus 4 and 128,000 for DeepSeek-R1. Maximum output per response: DeepSeek-R1 up to 32,768, GPT-4.1 up to 32,768, Claude Opus 4 up to 32,000 tokens.

Which can read images, PDFs, audio or video?

DeepSeek-R1 accepts text; GPT-4.1 accepts text, images and PDFs; Claude Opus 4 accepts text, images and PDFs. GPT-4.1 handles the widest range of inputs.

Are any of these open source?

DeepSeek-R1 publishes its weights and can be self-hosted; GPT-4.1 and Claude Opus 4 is proprietary.

Which is newer?

Claude Opus 4 is the newest, released May 22, 2025. GPT-4.1 came out Apr 14, 2025; DeepSeek-R1 came out Jan 20, 2025. Knowledge cutoff: DeepSeek-R1 Jul 2024, GPT-4.1 Apr 2024, Claude Opus 4 Mar 31, 2025.

How do you decide the winner?

This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 18, 2026.