ZMIME
Comparison · 3 models · Updated Nov 22, 2026

PaddleOCR VL vs GPT-5-Codex vs MinerU2.5 2509 1.2B

No independent capability benchmark covers every model here yet. The score uses price and context window.No independent capability benchmark covers every model here yet. The score uses context window.

  1. PaddlePaddle

    PaddleOCR VL

    Released Oct 16, 2025

    81/100
    • ECI—
    • Price$0.02 / $0.02
    • Context131K
  2. OpenAI

    GPT-5-Codex

    Released Sep 15, 2025

    31/100
    • ECI—
    • Price$1.25 / $10.00
    • Context400K
  3. Opendatalab

    MinerU2.5 2509 1.2B

    Released Sep 17, 2025

    71/100
    • ECI—
    • Price$0.024 / $0.071
    • Context16K
  4. PaddlePaddle

    PaddleOCR VL

    Released Oct 16, 2025

    32/100
    • ECI—
    • Price$0.02 / $0.02
    • Context131K
  5. OpenAI

    GPT-5-Codex

    Released Sep 15, 2025

    50/100
    • ECI—
    • Price$1.25 / $10.00
    • Context400K
  6. Opendatalab

    MinerU2.5 2509 1.2B

    Released Sep 17, 2025

    0/100
    • ECI—
    • Price$0.024 / $0.071
    • Context16K
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (PaddleOCR VL 81/100, MinerU2.5 2509 1.2B 71/100, GPT-5-Codex 31/100) compare the available specifications and price. The score weighs price 71%, context window 29%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.There is not enough shared capability evidence to name an overall winner. The scores (GPT-5-Codex 50/100, PaddleOCR VL 32/100, MinerU2.5 2509 1.2B 0/100) compare the available specifications only. The score weighs context window 100%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

  • CapabilityNot enough dataNo independent benchmark covers every model here yet
  • Lowest pricePaddleOCR VLPaddleOCR VL $0.02 · MinerU2.5 2509 1.2B $0.036 · GPT-5-Codex $3.44 per 1M tokens (3:1 blend)
  • Longest contextGPT-5-CodexGPT-5-Codex 400,000 · PaddleOCR VL 131,072 · MinerU2.5 2509 1.2B 16,384 tokens
  • Widest inputsSame inputsPaddleOCR VL: Text, Images · GPT-5-Codex: Text, Images · MinerU2.5 2509 1.2B: Text, Images
  • Self-hostingPaddleOCR VL and MinerU2.5 2509 1.2BPublishes downloadable weights (Apache-2.0 and AGPL-3.0)
How the score is built
MeasureWeightPaddleOCR VLGPT-5-CodexMinerU2.5 2509 1.2B
Price71%10023100
Context window29%32500
Specs & price only100%81/10031/10071/100

Left out because at least one model lacks the data: capability and inputs & features. The remaining weights were rescaled.

How the score is built
MeasureWeightPaddleOCR VLGPT-5-CodexMinerU2.5 2509 1.2B
Context window100%32500
Specs only100%32/10050/1000/100

Left out because at least one model lacks the data: capability and inputs & features. The remaining weights were rescaled.

02 — Side by side

Every spec in one table

Highlighted cells lead their row. Unavailable specifications are marked “Not published”. Publisher results include their evaluation settings and source.

PaddleOCR VL vs GPT-5-Codex vs MinerU2.5 2509 1.2B specifications side by side
SpecificationPaddleOCR VLPaddlePaddleGPT-5-CodexOpenAIMinerU2.5 2509 1.2BOpendatalab
Capability
Capabilities Index (ECI)Not publishedNot publishedNot published
ECI rankNot publishedNot publishedNot published
ECI confidence intervalNot publishedNot publishedNot published
ECI evaluation dateNot publishedNot publishedNot published
Documented capabilitiesCompact vision-language model for multilingual document parsing, OCR, tables, formulas, and chartsCoding-optimized GPT model for repository edits, reviews, and agentic software work1.2B-parameter vision-language model from OpenDataLab for two-step document parsing: global layout analysis on downsampled images, then fine-grained text, formula, and table recognition at native resolution
Model parametersNot publishedNot publishedNot published
Output modalitiesTextTextText
Publisher-reported benchmarks
Artificial Analysis Coding IndexPublisher-reported resultNot published38.9indexReported resultJun 1, 2026Not published
SciCodePublisher-reported resultNot published40.9percent correctReported resultJun 1, 2026Not published
Terminal-Bench HardPublisher-reported resultNot published37.9success rateReported resultJun 1, 2026Not published
Price per million tokens
Input$0.02 (best)$1.25$0.024
Output$0.02 (best)$10.00$0.071
Cached inputNot publishedNot publishedNot published
Blended (3:1)$0.02 (best)$3.44$0.036
Long-context rateSame rateSame rateSame rate
Price sourceMedian of 1 providersMedian of 9 providersMedian of 1 providers
Limits
Context window131,072 tokens400,000 tokens (best)16,384 tokens
Max inputNot published272,000 tokensNot published
Max output16,384 tokens128,000 tokens (best)Not published
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsNoNoNo
AudioNoNoNo
VideoNoNoNo
ReasoningNoYesNo
Tool callingNoYesNo
Structured outputNot publishedNot publishedNot published
File attachmentsYesYesYes
Temperature controlYesNoYes
Reasoning controlNot publishedalwaysNot published
Availability
WeightsOpenApache-2.0ProprietaryOpenAGPL-3.0
API model IDNot publishedNot publishedNot published
API providers110 (best)1
ReleasedOct 16, 2025Sep 15, 2025Sep 17, 2025
Knowledge cutoffNot publishedSep 30, 2024Not published
Publisher documentationModel weightsNot publishedModel weights
Source metadata updatedOct 16, 2025Sep 15, 2025Sep 17, 2025
Data sourcesmodels.dev · Epoch AImodels.dev · Epoch AImodels.dev · Epoch AI
Catalog checkedNov 22, 2026Nov 22, 2026Nov 22, 2026
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.

  • PaddleOCR VL$0.24
  • GPT-5-Codex$32.50
  • MinerU2.5 2509 1.2B$0.381
04 — Questions

Which should you choose?

Which is better: PaddleOCR VL, GPT-5-Codex or MinerU2.5 2509 1.2B?

There is not enough shared capability evidence to name an overall winner. The scores (PaddleOCR VL 81/100, MinerU2.5 2509 1.2B 71/100, GPT-5-Codex 31/100) compare the available specifications and price. The score weighs price 71%, context window 29%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.There is not enough shared capability evidence to name an overall winner. The scores (GPT-5-Codex 50/100, PaddleOCR VL 32/100, MinerU2.5 2509 1.2B 0/100) compare the available specifications only. The score weighs context window 100%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

Which is cheaper, PaddleOCR VL, GPT-5-Codex or MinerU2.5 2509 1.2B?

PaddleOCR VL is cheaper at $0.02 input / $0.02 output per million tokens (median across 1 API provider). MinerU2.5 2509 1.2B costs $0.024 input / $0.071 output per million tokens (median across 1 API provider); GPT-5-Codex costs $1.25 input / $10.00 output per million tokens (median across 9 API providers). At a typical mix of three input tokens to one output token, that is $0.02 per million tokens for PaddleOCR VL versus $0.036 for MinerU2.5 2509 1.2B (1.8× as much) and $3.44 for GPT-5-Codex (172× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. PaddleOCR VL has not been scored yet, GPT-5-Codex has not been scored yet and MinerU2.5 2509 1.2B has not been scored yet.

Which is better for coding?

There are no published SWE-bench Verified results for PaddleOCR VL, GPT-5-Codex and MinerU2.5 2509 1.2B yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. Note that PaddleOCR VL and MinerU2.5 2509 1.2B has no confirmed tool-calling support in this comparison.

Which has the bigger context window?

GPT-5-Codex has the largest context window at 400,000 tokens, against 131,072 for PaddleOCR VL and 16,384 for MinerU2.5 2509 1.2B. Maximum output per response: PaddleOCR VL up to 16,384, GPT-5-Codex up to 128,000 tokens.

Which can read images, PDFs, audio or video?

PaddleOCR VL accepts text and images; GPT-5-Codex accepts text and images; MinerU2.5 2509 1.2B accepts text and images. They handle the same number of input types.

Are any of these open source?

PaddleOCR VL and MinerU2.5 2509 1.2B publishes its weights (Apache-2.0 and AGPL-3.0) and can be self-hosted; GPT-5-Codex is proprietary.

Which is newer?

PaddleOCR VL is the newest, released Oct 16, 2025. MinerU2.5 2509 1.2B came out Sep 17, 2025; GPT-5-Codex came out Sep 15, 2025. Knowledge cutoff: GPT-5-Codex Sep 30, 2024.

How do you decide the winner?

This rating uses price (71%, blended price per million tokens (3 input : 1 output), log scale); context window (29%, 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 22, 2026.This rating uses context window (100%, 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 22, 2026.