Google's Bard, powered by the Gemini Pro-scale model, has surged to the second spot on the lmsys chatbot arena leaderboard, surpassing GPT-4. The achievement has sparked excitement and speculation about the potential release of Gemini Ultra. The performance of Bard with Gemini Pro has significantly improved, with users praising its capabilities for various tasks. The lower cost of Gemini Pro compared to GPT-4 turbo has also been highlighted. The competition between Google and OpenAI in the AI language model space is intensifying, with expectations for OpenAI to release a new model to maintain its lead. The comparison between Bard and GPT-4 has generated discussions about the nuances of evaluating language models and the impact of different APIs and models on performance.
Very interesting article by @MikeKaput over at @MktgAi ๐ Google Bard Now Surpasses GPT-4: https://t.co/XIn28cZ5Wy Bard is now powered by Gemini Pro, and since adding it to @HeyMagai last month, I can honestly say it's now my most used model for content writing. And I loveโฆ
Instead of renaming Bard to Gemini, Google should just rename it Google.
and this is only Gemini Pro that's beating GPT4-V, just wait for Ultra https://t.co/6HnaVrJRpW https://t.co/XHeVomFpMd
So looks like the new GPT4 Turbo has more updated data. https://t.co/OmPrvpJ6KS
Googleโs Assistant with Bard might rebrand to Gemini https://t.co/uxI3vaqrFk https://t.co/RhufSLV8hS
Google may reconsider the clunky โAssistant with Bardโ name ahead of launch https://t.co/LIwfxIG5DW
๐จ Google Bard is coming for ChatGPT's crown. The search giant's Gemini Pro model has climbed to second spot on the HuggingFace Chatbot Arena leaderboard. It sits only behind GPT-4 Turbo, but has beaten the regular ol' GPT-4. https://t.co/Gfz5tTs9Fv
Really enjoying Bard all of a sudden. Huge improvement compared to a few weeks ago. Fast, mostly accurate, and pretty smart. Looking forward to Gemini Ultra/Bard Pro and OpenAI's inevitable response.
New Gemini pro update should be hitting bard Feb 1st Might be the one that beat gpt-4 in the arena Competition is heating up. Wen gpt5 https://t.co/kLCpPcjsOV
Google might be considering another name for Assistant with Bard: Gemini https://t.co/6ZTbyhGeIO by @technacity
Early look at Googleโs Bard-powered Assistant emerges online https://t.co/WG97Yym3zb https://t.co/SlEUILDYKm
Google plans to add a Bard Assistant in the next Pixel feature drop https://t.co/SA5L7QTe53 https://t.co/9xd9NGLc6u
The excitement around the Bard leaderboard release has been incredible! A few things to note: 1. Arena eval is LIVE as we tweet. If in doubt, we encourage everyone to compare models firsthand in Arena and contribute your use case and vote๐ณ๏ธ Let's make this community-drivenโฆ https://t.co/IgXzbcdRw0
An interesting thing about this result is how much better "Bard (Gemini Pro)" is than "Gemini Pro". It emphasizes the importance of good wrappers *around* strong LMs. It would be really interesting to se "ChatGPT (gpt-3.5-turbo)" and "ChatGPT (gpt-4-turbo)" on the ranking as well https://t.co/blAvVyDMRR
Bard (Gemini Pro) is living in your head rent free.
Is that true, though? Gemini Pro dev API, which costs 50 cents, still significantly underperforms GPT-4, and the result everyone is talking about is for โBard with Gemini Proโ. There is not even an API available for that, let alone pricing. https://t.co/J1hTgkeLpS
๐๏ธ Today in AI (Jan 27th, 2024) - 5 Minute TLDR - Google has released Bard Gemini Ultra, which is on par with GPT-4 in human evaluation. Oriol Vinyals, head of deep learning at Google and co-lead of Gemini, highlighted the challenges of evaluating language models and theโฆ https://t.co/fxkWLKU0yH
Many seem confused by this and correctly so. Bard, unlike the Gemini pro dev API, has access to the web and possibly has additional stuff going into it to make it good enough for an average user. This leaderboard is not an apples-to-apples comparison, as we donโt know ifโฆ https://t.co/GOjnsOoyqW
Bard with Gemini Pro has surpassed GPT-4 on Arena. Combine it with Google Sheets to automate your data processing. I show you how to use Bard to manage your spreadsheets without formulas: https://t.co/eEw5pKpWyw
Gemini Pro is now ranking higher than GPT-4. This has shocked many users. So I decided to compare both on: - Critical thinking - Math - Programming - Creative writing Here are the results ๐งต https://t.co/1MZyed10Tg
Apparently a new version of GPT-4 dropped and it might be much better at reasoning. https://t.co/oklh4pluks
Google Bard Gemini Pro beats GPT-4 to become the second-best LLM. Remember, we're still waiting for Gemini Ultra. So, is OpenAI's new model soon? I have tested the updated Gemini Pro with ChatGPT-4: https://t.co/TzLdEKU9YY
Ahhhh, so that's why I said: there's no way Gemini beats older GPT-4 or Claudeโthere's no way. Have you actually tried using it?! Turns out: there are actually three versions of Gemini on this table. Gemini via API ranks below Claude 2.0โthat was obvious. Time to try Bard! https://t.co/AD7MSMtaPT
Google's Bard has just made a stunning leap, surpassing GPT-4 to the SECOND SPOT on the leaderboard! https://t.co/uIWGpdmRjC #ArtificialIntelligence @GoogleAI
Everyone's talking about Gemini outperforming older GPT-4 models on benchmarks โbut are you actually USING Gemini over GPT-4 for your work? I ran 1k+ API calls to both models last week and, for my task, Gemini wasn't even close. How does it do for your work? https://t.co/H0ufb9V481
I believe itโs called Bard with Gemini Pro https://t.co/k8T0ZUwQXe
I love Chatbot Arena, but this is crazy. "GPT-4-Turbo" is a bunch of model weights. "Bard (Gemini Pro)" is the Google crawler/index + Gemini There's no way on earth a live crawler should be compared to base model weights. If anything, it's actually crazy that GPT-4 disconnectedโฆ https://t.co/7Xr3pWxgaU
I love Chatbot Arena, but this is crazy. "GPT-4-Turbo" is a bunch of model weights. "Bard (Gemini Pro)" is the Google crawler/index + Bard There's no way on earth a live crawler should be compared to base model weights. If anything, it's actually crazy that GPT-4 disconnectedโฆ https://t.co/7Xr3pWxgaU
I challenge everyone who is sharing that Bard/Gemini is reaching the same level as GPT-4 to test Bard, Gemini. Real testing. Difficult problems. Stop being influenced by tweets. The competition is still far away. Very far away.
Googles Face when Bard jumps the benchmarks https://t.co/gk9OZ3xndc https://t.co/NvtsV53Dty
It's crazy how people are influenced by tweets, especially from influencers and benchmarks. Gemini/Bard has always been mocked for being bad, everyone who tests it maintains the same opinion that still bad, but now I see a lot Bardguy because it scored well in the chatarena.
Evaluation of LLMs is very hard and nuanced (especially academic evals which are leaked massively). Evals that rely on human judgement are far superior, so it feels good that Bard Gemini Pro (free tier) climbed pretty high on lmsys๐ช Looking forward to Gemini Ultra release!โฆ https://t.co/3ngJ217rCW
i don't put much stock in LLM leaderboards, but if you haven't checked out Bard lately you really should the past couple months it's been absolutely fantastic for creative writing tasks https://t.co/esqHFIJy34
We probably should put Bard Gemini Nano-2 on lmsys and show what that model can do, right?
Is bard just retrieval augumentation + geminiPro? If so has GPT4 been the same this whole time? The I wonder how would perplexityโs private api perform on LMsys arena benchmark too.
Wow! Bard has surpassed GPT-4. This is a leaderboard I follow and trust. "LMSYS Chatbot Arena Leaderboard" Link to the HuggingFace leaderboard in the post. https://t.co/8iiZuKalia
๐ฅLooks like Google released a new model and it's #2 on the leaderboards, beating out some GPT-4 models?! "bard-jan-24-gemini-pro" https://t.co/6zUOWwTtu6
Google Bard Gemini Pro has entered the Arena ๐คฏ https://t.co/Fwqt0vdIGJ
It looks like Google has significantly improved bard, thanks to Gemini Pro's latest update. Note that Gemini Pro is Google's second best model. The best one, not released yet, is Gemini Ultra. See link below to know everything about Gemini models https://t.co/IH79XRkF8N
Gemini Pro matches GPT-4 in ELO, but the real game changer is that sampling a million tokens costs $0.5 instead of $30 https://t.co/UrfnVvKzmF
Note that this is the Gemini Pro Model. The best is yet to come โค๏ธ! Gemini Ultra getting ready ๐ฅ๐ฅ๐ฅ! https://t.co/0XYFT3wxqZ
Google made massive improvements to Bard, with Gemini Pro now ranking higher than regular GPT-4. Keep in mind, Gemini Pro is Google's SECOND best model. We're still waiting on Gemini Ultra. Expect OpenAI to drop a new model anytime now to stay in the lead. https://t.co/AlM3JVQraB
This is amazing if Gemini Pro is a GPT-3.5 scale model. https://t.co/eyHfnn5uhi
By the time Gemini ultra is available to plebs, gpt 5 will be broadly available https://t.co/Jezo7sDvYh
Bard (free) with Gemini Pro is now #2 on lmsys chatbot arena, surpassing GPT-4. I use it a lot for email and writing purposes and its responses are much better since the first launch. Give it a try at https://t.co/OxsIZaS0K7 https://t.co/Ey1GHdOm1E
Beware of Gemini! This just happened: Gemini Pro overcame every other GPT-4 version and it's now sitting at the 2nd spot in the LLM race. This is the Pro version. We are still waiting to see what Gemini Ultra can do. OpenAI's lead is almost gone. Expect their new model soon! https://t.co/6LBCuL4Gjj
Interestingly, Bard is better than the Gemini Pro API on the Elo rating! Is Bard powered by Ultra, today? ๐คฏ๐ Finally, after 20 months, a real challenger to GPT-4 may have emerged https://t.co/umkAs8DW3l
Bard powered by gemini pro is in the 1200 elo club on lmsys ๐ฅ It's a great model, and it's free! https://t.co/WiE99hhgx0
the https://t.co/i3Nj6uwTws leaderboard is my #1 go-to for LLM performance comparisons, and this is a big update! https://t.co/iaJ4w29Ny4
Go Google Bard! https://t.co/Mkf0mQlsaI
Bard (free) with Gemini Pro is now #2 on lmsys chatbot arena, surpassing GPT-4. I use it a lot for email purposes and its responses are much better since the first launch. Give it a try at https://t.co/OxsIZaS0K7 https://t.co/Ey1GHdOm1E
I now use Bard more than Chatgpt. It especially shines for new material. E.g. Bard really shines for queries needing latest results like "give me references discussing instruction tuning vs fine tuning". https://t.co/ikWEcjxZtU
Someone on my feed last night was saying Bard was suddenly better at code than GPT-4. Google is doing some work over there apparently. https://t.co/A1NSLaNEJU
A truly defining moment during the AI Spring. One reason this is a bigger deal is partly because Gemini Pro is *far* lower in price than GPT-4 turbo ~40x lower? https://t.co/bSsGT17kzc
Bard (free) with Gemini Pro is now #2 on lmsys chatbot arena. Give it a try at https://t.co/OxsIZaS0K7 https://t.co/Ey1GHdOm1E
Someone on my feed last night was saying Bard was suddenly better at code than GPT-4. Google is doing some work over there apparently. Possible pre-release stealth testing of Ultra behind the mask of Bard? https://t.co/A1NSLaNEJU
A truly defining moment during the AI Spring. One reason this is a bigger deal is partly because Gemini Pro is *far* lower in price than GPT-4 turbo. https://t.co/oM5JJERYZr
Bard, powered by the Gemini Pro-scale model, debuts at the #2 position on the independent lmsys leaderboard. ๐ฅ Give it a try at https://t.co/m9D7JYUfls. Bard is much better & has many more capabilities since its debut in March, thanks to everyone on the Bard/Gemini teams! https://t.co/rOPTbOE4v8
So exciting to see third party evaluation results on Bard! (free tier available today) Letโs just say its performance has been greatly underestimated.๐ค https://t.co/QRjJnH9Wot
Super excited that Bard is now doing so well on Chatbot Arena! https://t.co/y1hBSIvBtj
I wonder how Bard Advanced will stack upโฆ https://t.co/nkZVnTlBnh
๐ฅBreaking News from Arena Google's Bard has just made a stunning leap, surpassing GPT-4 to the SECOND SPOT on the leaderboard! Big congrats to @Google for the remarkable achievement! The race is heating up like never before! Super excited to see what's next for Bard + Geminiโฆ https://t.co/QPtsqZdJhC