Today, we’re releasing GPT‑5.4 in ChatGPT (as GPT‑5.4 Thinking), the API, and Codex. It’s our most capable and efficient frontier model for professional work. We’re also releasing GPT‑5.4 Pro in ChatGPT and the API, for people who want maximum performance on complex tasks.
GPT‑5.4 brings together the best of our recent advances in reasoning, coding, and agentic workflows into a single frontier model. It incorporates the industry-leading coding capabilities of GPT‑5.3‑Codex while improving how the model works across tools, software environments, and professional tasks involving spreadsheets, presentations, and documents. The result is a model that gets complex real work done accurately, effectively, and efficiently—delivering what you asked for with less back and forth.
In ChatGPT, GPT‑5.4 Thinking can now provide an upfront plan of its thinking, so you can adjust course mid-response while it’s working, and arrive at a final output that’s more closely aligned with what you need without additional turns. GPT‑5.4 Thinking also improves deep web research, particularly for highly specific queries, while better maintaining context for questions that require longer thinking. Together, these improvements mean higher-quality answers that arrive faster and stay relevant to the task at hand.
In Codex and the API, GPT‑5.4 is the first general-purpose model we’ve released with native, state-of-the-art computer-use capabilities, enabling agents to operate computers and carry out complex workflows across applications. It supports up to 1M tokens of context, allowing agents to plan, execute, and verify tasks across long horizons. GPT‑5.4 also improves how models work across large ecosystems of tools and connectors with tool search, helping agents find and use the right tools more efficiently without sacrificing intelligence. Finally, GPT‑5.4 is our most token efficient reasoning model yet, using significantly fewer tokens to solve problems when compared to GPT‑5.2—translating to reduced token usage and faster speeds.
Together with advances in general reasoning, coding, and professional knowledge work, GPT‑5.4 enables more reliable agents, faster developer workflows, and higher-quality outputs across ChatGPT, the API, and Codex.
GPT‑5.4 | GPT‑5.3‑Codex | GPT‑5.2 | |
GDPval (wins or ties) | 83.0% | 70.9% | 70.9% |
SWE-Bench Pro (Public) | 57.7% | 56.8% | 55.6% |
OSWorld-Verified | 75.0% | 74.0%* | 47.3% |
Toolathlon | 54.6% | 51.9% | 46.3% |
BrowseComp | 82.7% | 77.3% | 65.8% |
*Previously reported as 64.7%. GPT‑5.3‑Codex achieves 74.0% with a newly introduced API parameter that preserves the original image resolution.
Building on GPT‑5.2’s general reasoning capabilities, GPT‑5.4 delivers even more consistent and polished results on real-world tasks that matter to professionals.
On GDPval, which tests agents’ abilities to produce well-specified knowledge work across 44 occupations, GPT‑5.4 achieves a new state of the art, matching or exceeding industry professionals in 83.0% of comparisons, compared to 70.9% for GPT‑5.2.
U GDPval-u modeli pokušavaju da obave jasno definisan rad sa znanjem koji obuhvata 44 zanimanja iz 9 najvećih industrija koje doprinose BDP-u SAD. Zadaci traže stvarne radne rezultate, kao što su prodajne prezentacije, računovodstvene tabele, rasporedi hitne nege, proizvodni dijagrami ili kratki video-snimci. Nivo rezonovanja bio je podešen na xhigh za GPT‑5.4 i heavy za GPT‑5.2 (nešto niži nivo u ChatGPT‑u).
„GPT-5.4 je najbolji model koji smo ikada probali. Sada je na vrhu rang-liste na našem APEX-Agents benchmarku, koji meri performanse modela za rad u profesionalnim uslugama. Odličan je u kreiranju dugoročnih isporuka kao što su prezentacije, finansijski modeli i pravne analize, pružajući vrhunske performanse uz veću brzinu i niže troškove od konkurentskih graničnih modela.“
We put a particular focus on improving GPT‑5.4’s ability to create and edit spreadsheets, presentations, and documents. On an internal benchmark of spreadsheet modeling tasks that a junior investment banking analyst might do, GPT‑5.4 achieves a mean score of 87.3%, compared to 68.4% for GPT‑5.2. On a set of presentation evaluation prompts, human raters preferred presentations from GPT‑5.4 68.0% of the time over those from GPT‑5.2 due to stronger aesthetics, greater visual variety, and more effective use of image generation.
Dokumenti su generisani sa nivoom rezonovanja podešenim na xhigh
You can try these capabilities in ChatGPT using GPT‑5.4 Thinking or Pro. If you’re an Enterprise customer, we recommend using our newly released ChatGPT for Excel add-in(отвара се у новом прозору), which was also launched today. We've also updated our spreadsheet(отвара се у новом прозору) and presentation skills(отвара се у новом прозору) available in Codex and the API.
To make GPT‑5.4 better at real-world work, we continued our progress at driving down hallucinations and errors. GPT‑5.4 is our most factual model yet: on a set of de-identified prompts where users flagged factual errors, GPT‑5.4’s individual claims are 33% less likely to be false and its full responses are 18% less likely to contain any errors, relative to GPT‑5.2.
„GPT-5.4 postavlja novi standard za pravni rad sa velikim brojem dokumenata. Na našoj BigLaw Bench proceni postigao je 91%. U poređenju sa drugim modelima, GPT-5.4 je trenutno bolji u strukturiranju složene transakcione analize, održavanju tačnosti kroz dugačke ugovore i isporuci visokog nivoa detalja koji pravni stručnjaci zahtevaju.“
GPT‑5.4 is our first general-purpose model with native computer-use capabilities and marks a major step forward for developers and agents alike. It’s the best model currently available for developers building agents that complete real tasks across websites and software systems.
We’ve designed GPT‑5.4 to be performant across a wide range of computer-use workloads. It is excellent at writing code to operate computers via libraries like Playwright, as well as issuing mouse and keyboard commands in response to screenshots. Its behavior is steerable via developer messages, meaning that developers can adjust behavior to suit particular use cases. Developers can even configure the model’s safety behavior to suit different levels of risk tolerance by specifying custom confirmation policies.
The model’s performance and flexibility are reflected across benchmarks that test computer use across different settings. On OSWorld-Verified, which measures a model’s ability to navigate a desktop environment through screenshots and keyboard/mouse actions, GPT‑5.4 achieves a state-of-the-art 75.0% success rate, far exceeding GPT‑5.2’s 47.3%, and surpassing human performance at 72.4%.1
On WebArena-Verified, which tests browser use, GPT‑5.4 achieves a leading 67.3% success rate when using both DOM- and screenshot-driven interaction, compared to GPT‑5.2’s 65.4%. On Online-Mind2Web, which also tests browser use, GPT‑5.4 achieves a 92.8% success rate using screenshot-based observations alone, improving over ChatGPT Atlas’s Agent Mode, which achieves a success rate of 70.9%.
Prepuštanje alatu nastaje kada asistent prepusti obradu čekajući odgovore alata. Ako se 3 alata pozovu paralelno, a zatim još 3 alata paralelno, broj prepuštanja bio bi 2. Prepuštanja alatima su bolji pokazatelj latencije od poziva alata jer odražavaju prednosti paralelizacije.
GPT‑5.4 tumači snimke ekrana interfejsa pregledača i stupa u interakciju sa UI elementima klikovima zasnovanim na koordinatama kako bi slao imejlove i zakazivao događaj u kalendaru. Video nije ubrzan.
GPT‑5.4’s improved computer use is built on the model’s improved general visual perception capabilities. On MMMU-Pro, a test of a model’s visual understanding and reasoning, GPT‑5.4 achieves an 81.2% success rate without tool use, an improvement over GPT‑5.2’s 79.5%. Improved visual perception also translates into better document parsing capabilities. On OmniDocBench, GPT‑5.4 without reasoning effort achieves an average error (measured by normalized edit distance between model prediction and ground truth) of 0.109, improved from GPT‑5.2’s 0.140.
MMMUPro je pokrenut sa nivoom rezonovanja podešenim na xhigh. OmniDocBench je pokrenut sa nivoom rezonovanja podešenim na none, kako bi odražavao performanse sa niskim troškom i niskom latencijom.
We’re also improving visual understanding for dense, high-resolution images where full fidelity matters. Starting with GPT‑5.4, we’re introducing an original image input detail(отвара се у новом прозору) level which supports full-fidelity perception up to 10.24M total pixels or 6000-pixel maximum dimension, whichever is lower; the high image input detail level now supports up to 2.56M total pixels or a 2048-pixel maximum dimension. In early testing with API users, we observed strong gains in localization ability, image understanding, and click accuracy when using original or high detail.
„U našim procenama koje mere performanse korišćenja računara na ~30K HOA i portalima za porez na imovinu, GPT-5.4 je postigao 95% uspešnosti iz prvog pokušaja i 100% u okviru tri pokušaja, u poređenju sa ~73–79% kod prethodnih CUA modela. Takođe je završavao sesije oko ~3x brže uz ~70% manje tokena, što značajno poboljšava pouzdanost i troškovnu efikasnost u velikom obimu."
In the API, developers can access these capabilities using the updated computer tool. Please see our updated documentation(отвара се у новом прозору) for recommended best practices.
GPT‑5.4 combines the coding strengths of GPT‑5.3‑Codex with leading knowledge work and computer-use capabilities, which matter most on longer-running tasks where the model can use tools, iterate, and push work further with less manual intervention. It matches or outperforms GPT‑5.3‑Codex on SWE-Bench Pro while being lower latency across reasoning efforts.
Latenciju procenjujemo posmatranjem ponašanja naših modela u produkciji i simuliranjem toga van mreže. Procena latencije uzima u obzir trajanje poziva alata (vreme izvršavanja koda), uzorkovane tokene i ulazne tokene. Latencija u stvarnom svetu može znatno da varira i zavisi od mnogo faktora koji nisu obuhvaćeni našom simulacijom. Nivoi rezonovanja kretali su se od none do xhigh.
When toggled on, /fast mode in Codex delivers up to 1.5x faster token velocity with GPT‑5.4. It’s the same model and the same intelligence, just faster. That means users can move through coding tasks, iteration, and debugging while staying in flow. Developers can access GPT‑5.4 at the same fast speeds via the API by using priority processing(отвара се у новом прозору).
In evaluation and internal testing we found that GPT‑5.4 excels at complex frontend tasks, with noticeably more aesthetic and more functional results than any models we’ve launched previously.
As a demonstration of the model’s improved computer-use and coding capabilities working in tandem, we’re also releasing an experimental Codex skill called “Playwright (Interactive)(отвара се у новом прозору)”. This allows Codex to visually debug web and Electron apps; it can even be used to test an app it’s building, as it’s building it.
Simulaciona igra zabavnog parka napravljena pomoću GPT‑5.4 iz jedne minimalno specificirane instrukcije, uz korišćenje Playwright Interactive za testiranje u pregledaču i generisanje slika za izometrijski skup resursa. Simulacija uključuje postavljanje staza po pločicama, izgradnju vožnji i scenografije, pronalaženje putanje gostiju, čekanje u redu i cikluse vožnji, dok metrike parka poput novca, broja gostiju, zadovoljstva, čistoće i ocene rastu ili padaju u zavisnosti od toga kako raspored funkcioniše i kako gosti reaguju na njega. Playwright je korišćen za automatizaciju testiranja u pregledaču kroz izgradnju i proširivanje parka, postavljanje i uklanjanje staza i atrakcija, proveru navigacije kamere i verifikaciju da su se gosti, redovi, stanja vožnji i UI metrike ispravno ažurirali tokom više rundi igre.
Instrukcija: Use $playwright-interactive and $imagegen. Create an interactive isometric theme park simulation game that I can build and navigate in the browser. Use imagegen to establish the overall visual vision and generate the game's assets, including rides, paths, terrain, trees, water, food stalls, decorations, buildings, icons, and UI illustrations. The world should feel cohesive, polished, and visually rich, with a premium art direction that works well from an isometric perspective. Let me place and remove paths, add attractions, position scenery, and move around the park smoothly while monitoring guest activity, ride status, and park growth. Include believable guest movement, simple park management systems like money, cleanliness, queueing, and happiness, and make the experience feel playful, clear, and complete rather than like a rough prototype. Prioritize charm, readability, and strong game feel over realism.
When play testing, be sure to build and expand a park through several rounds of play, verify that placement and navigation work smoothly, confirm that guests react to the park layout and attractions, and ensure the visuals, UI, and interactions feel stable and cohesive.
„GPT-5.4 je trenutno vodeći na našim internim benchmarkovima. Naši inženjeri smatraju da je prirodniji i odlučniji od prethodnih modela. Prolazi kroz dvosmislene probleme bez preispitivanja samog sebe i proaktivan je u paralelizaciji rada kako bi se stvari odvijale bez zastoja.“
With GPT‑5.4, we’ve significantly improved how models work with external tools. Agents can now operate across larger tool ecosystems, choose the right tools more reliably, and complete multi-step workflows with lower cost and latency.
In the API, GPT‑5.4 introduces tool search(отвара се у новом прозору), which allows models to work efficiently when given many tools.
Previously, when a model was given tools, all tool definitions were included in the prompt upfront. For systems with many tools, this could add thousands—or even tens of thousands—of tokens to every request, increasing cost, slowing responses, and crowding the context with information the model might never use.
With tool search, GPT‑5.4 instead receives a lightweight list of available tools along with a tool search capability. When the model needs to use a tool, it can look up that tool’s definition and append it to the conversation at that moment.
This approach dramatically reduces the number of tokens required for tool-heavy workflows and preserves the cache, making requests faster and cheaper. It also enables agents to reliably work with much larger tool ecosystems. For MCP servers that may contain tens of thousands of tokens of tool definitions, the efficiency gains can be substantial.
To demonstrate the efficiency gains, we evaluated 250 tasks from Scale’s MCP Atlas(отвара се у новом прозору) benchmark with all 36 MCP servers enabled in two modes: (1) exposing every MCP function directly in the model context, and (2) placing all MCP servers behind tool search. The tool-search configuration reduced total token usage by 47% while achieving the same accuracy.
Primeri broja tokena dobijeni su prosekom preko 250 zadataka iz javnog skupa podataka MCP-Atlas.
GPT‑5.4 also improves tool calling, making it more accurate and efficient when deciding when and how to use tools during reasoning, particularly in the API. Compared to GPT‑5.2, it achieves higher accuracy in fewer turns on Toolathlon, a benchmark that tests how well AI agents can use real-world tools and APIs to complete multi-step tasks. For example, an agent needs to read emails, extract assignment attachments, upload them, grade them and record results in a spreadsheet.
Prepuštanje alatu nastaje kada asistent prepusti obradu čekajući odgovore alata. Ako se 3 alata pozovu paralelno, a zatim još 3 alata paralelno, broj prepuštanja bio bi 2. Prepuštanja alatima su bolji pokazatelj latencije od poziva alata jer odražavaju prednosti paralelizacije.
For latency-sensitive use cases where reasoning effort None is preferred, GPT‑5.4 further improves upon its predecessors.
U τ2-bench(отвара се у новом прозору)-u, model mora da koristi alate kako bi obavio zadatak korisničke podrške, pri čemu može postojati simulirani korisnik koji može da komunicira i preduzima radnje nad stanjem sveta. Nivo rezonovanja bio je podešen na None.
GPT‑5.4 is better at agentic web search. On BrowseComp, a measurement of how well AI agents can persistently browse the web to find hard-to-locate information, GPT‑5.4 leaps 17%abs over GPT‑5.2, and GPT‑5.4 Pro sets a new state of the art of 89.3%.
In practice, this means GPT‑5.4 Thinking is stronger at answering questions that require pulling together information from many sources on the web. It can more persistently search across multiple rounds to identify the most relevant sources, particularly for “needle-in-a-haystack” questions, and synthesize them into a clear, well-reasoned answer.
U BrowseComp-u koristili smo blocklist-u pretrage koja iz procene isključuje veb-sajtove sa odgovorima iz benchmarka kako bismo sprečili kontaminaciju i obezbedili fer meru performansi. GPT‑5.4 je meren kasnije od GPT‑5.2, pa rezultati odražavaju promene u modelu, našem sistemu pretrage i stanju interneta. GPT‑5.4 je testiran sa dužom, ažuriranom blocklist-om. Modeli koriste ChatGPT alat za pretragu, koji može imati male razlike u odnosu na API pretragu.
„GPT-5.4 xhigh je novi vrhunac tehnike za višekoračnu upotrebu alata. Zapier sprovodi neka od najrigoroznijih merenja upotrebe alata u industriji, testirajući modele kroz stotine naprednih tokova rada iz stvarnog sveta. GPT-5.4 je završio posao tamo gde su prethodni modeli odustajali — najuporniji model do sada.“
Similarly to how Codex outlines its approach when it starts working, GPT‑5.4 Thinking in ChatGPT will now outline its work with a preamble for longer, more complex queries. You can also add instructions or adjust its direction mid-response. This makes it easier to guide the model toward the exact outcome you want without starting over or requiring multiple additional turns. This feature is available now on chatgpt.com(отвара се у новом прозору) and the Android app, coming soon to the iOS app.
The model can also think longer on difficult tasks while maintaining stronger awareness of earlier steps in the conversation. This allows it to handle longer workflows and more complex prompts while keeping answers coherent and relevant throughout.
Ovaj video je ubrzan u ilustrativne svrhe.
Over recent months, we’ve continued improving the safeguards we introduced with GPT‑5.3‑Codex while preparing GPT‑5.4 for deployment. Similar to GPT‑5.3‑Codex, we are treating GPT‑5.4 as High cyber capability under our Preparedness Framework, and we are deploying it with the corresponding protections as documented in the system card(отвара се у новом прозору). These include an expanded cyber safety stack, including monitoring systems, trusted access controls, and asynchronous blocking for higher-risk requests for customers on Zero Data Retention (ZDR) surfaces, alongside ongoing investment in the broader security ecosystem.
Because cybersecurity capabilities are inherently dual-use, we maintain a precautionary approach to deployment while continuing to calibrate our policies and classifiers. For certain customers on ZDR surfaces, request-level blocking remains part of our cyber risk mitigation stack; because classifiers are still improving, some false positives may occur as we continue refining these safeguards. These updates aim to improve how safeguards operate in practice, including reducing unnecessary refusals and overly caveated responses, while preserving strong protections against misuse.
We’ve continued our safety research on Chain-of-Thought (CoT) monitorability to better understand how models reason and help detect potential misbehavior. As part of this work, we introduce a new open-source evaluation, CoT controllability, measuring whether models can deliberately obfuscate their reasoning to evade monitoring. We find that GPT‑5.4 Thinking’s ability to control its CoT is low, which is a positive property for safety, suggesting that the model lacks the ability to hide its reasoning and that CoT monitoring remains an effective safety tool.
GPT‑5.4 is rolling out gradually today across ChatGPT and Codex. In the API, GPT‑5.4 is available now as gpt-5.4. GPT‑5.4 Pro is also available in the API as gpt-5.4-pro for developers who need maximum performance on the most complex tasks.
In ChatGPT, GPT‑5.4 Thinking is available starting today to ChatGPT Plus, Team, and Pro users, replacing GPT‑5.2 Thinking. GPT‑5.2 Thinking will remain available for three months for paid users in the model picker under the Legacy Models section, after which it will be retired on June 5, 2026. Those on Enterprise and Edu plans can enable early access via admin settings. GPT‑5.4 Pro is available to Pro and Enterprise plans. Context windows(отвара се у новом прозору) in ChatGPT for GPT‑5.4 Thinking remain unchanged from GPT‑5.2 Thinking.
GPT‑5.4 is our first mainline reasoning model that incorporates the frontier coding capabilities of GPT‑5.3‑codex and that is rolling out across ChatGPT, the API and Codex. We're calling it GPT‑5.4 to reflect that jump, and to simplify the choice between models when using Codex. Over time, you can expect our Instant models and Thinking models to evolve at different speeds.
GPT‑5.4 in Codex includes experimental support for the 1M context window. Developers can try this by configuring model_context_window and model_auto_compact_token_limit. Requests that exceed the standard 272K context window count against usage limits at 2x the normal rate.
In the API, GPT‑5.4 is priced higher per token than GPT‑5.2 to reflect its improved capabilities, while its greater token efficiency helps reduce the total number of tokens required for many tasks. Batch and Flex pricing are available at half the standard API rate, while Priority processing is available at twice the standard API rate.
API model | Cena ulaza | Cena keširanog ulaza | Cena izlaza |
gpt-5.2 | $1.75 / M tokena | $0.175 / M tokena | $14 / M tokena |
gpt-5.4 | $2.50 / M tokena | $0.25 / M tokena | $15 / M tokena |
gpt-5.2-pro | $21 / M tokena | - | $168 / M tokena |
gpt-5.4-pro | $30 / M tokena | - | $180 / M tokena |
Profesionalno
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
GDPval | 83.0% | 82.0% | 70.9% | 70.9% | 74.1% |
FinanceAgent v1.1 | 56.0% | 61.5% | 54.0% | 59.5% | — |
Zadaci modelovanja za investiciono bankarstvo (interno) | 87.3% | 83.6% | 79.3% | 68.4% | 71.7% |
OfficeQA | 68.1% | — | 65.1% | 63.1% | — |
Kodiranje
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
SWE-Bench Pro (javno) | 57.7% | — | 56.8% | 55.6% | — |
Terminal-Bench 2.0 | 75.1% | — | 77.3% | 62.2% | — |
Korišćenje računara i vid
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
OSWorld-Verified | 75.0% | — | 74.0% | 47.3% | — |
MMMU Pro (bez alata) | 81.2% | — | — | 79.5% | — |
MMMU Pro (sa alatima) | 82.1% | — | — | 80.4% | — |
Upotreba alata
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
BrowseComp | 82.7% | 89.3% | 77.3% | 65.8% | 77.9% |
MCP Atlas | 67.2% | — | — | 60.6% | — |
Toolathlon | 54.6% | — | 51.9% | 45.7% | — |
Tau2-bench Telecom | 98.9% | — | — | 98.7% | — |
Akademsko
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
Frontier Science Research | 33.0% | 36.7% | — | 25.2% | — |
FrontierMath Tier 1–3 | 47.6% | 50.0% | — | 40.7% | — |
FrontierMath Tier 4 | 27.1% | 38.0% | — | 18.8% | 31.3% |
GPQA Diamond | 92.8% | 94.4% | 92.6% | 92.4% | 93.2% |
Humanity's Last Exam (bez alata) | 39.8% | 42.7% | — | 34.5% | 36.6% |
Humanity's Last Exam (sa alatima) | 52.1% | 58.7% | — | 45.5% | 50.0% |
Dugačak kontekst
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
Graphwalks BFS 0K–128K | 93.0% | — | — | 94.0% | — |
Graphwalks BFS 256K–1M | 21.4% | — | — | — | — |
Graphwalks parents 0–128K (tačnost) | 89.8% | — | — | 89.0% | — |
Graphwalks parents 256K–1M (tačnost) | 32.4% | — | — | — | — |
OpenAI MRCR v2 8-needle 4K–8K | 97.3% | — | — | 98.2% | — |
OpenAI MRCR v2 8-needle 8K–16K | 91.4% | — | — | 89.3% | — |
OpenAI MRCR v2 8-needle 16K–32K | 97.2% | — | — | 95.3% | — |
OpenAI MRCR v2 8-needle 32K–64K | 90.5% | — | — | 92.0% | — |
OpenAI MRCR v2 8-needle 64K–128K | 86.0% | — | — | 85.6% | — |
OpenAI MRCR v2 8-needle 128K–256K | 79.3% | — | — | 77.0% | — |
OpenAI MRCR v2 8-needle 256K–512K | 57.5% | — | — | — | — |
OpenAI MRCR v2 8-needle 512K–1M | 36.6% | — | — | — | — |
Apstraktno rezonovanje
Procena | GPT‑5.4 | GPT‑5.4 | GPT‑5.3-Codex | GPT‑5.2 | GPT‑5.2 |
ARC-AGI-1 (verifikovano) | 93.7% | 94.5% | — | 86.2% | 90.5% |
ARC-AGI-2 (verifikovano) | 73.3% | 83.3% | — | 52.9% | 54.2% (high) |
Procene bez rezonovanja
Procena | GPT‑5.4 | GPT‑5.2 | GPT‑4.1 |
OmniDocBench (normalizovana udaljenost uređivanja) | 0.109 | 0.140 | — |
Tau2-bench Telecom | 64.3% | 57.2% | 43.6% |
Evals were run with reasoning effort set to xhigh, except where specified otherwise. Benchmarks were conducted in a research environment, which may provide slightly different output from production ChatGPT in some cases.
Аутор
Fusnote
1 Ljudske performanse navedene u radu OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments(отвара се у новом прозору).
Наставите са читањем
