Arena.ai(@lmarena_ai)

Dive into the Agent Arena at: https://t.co/sE9q4FSYAt

Arena.ai (@arena) — Mon, 19 Oct 2026 18:18:51 GMT

Dive into the Agent Arena at: arena.ai/leaderboard/ag…

💬0🔄1❤️4👀3312📊1

Big news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Agent Arena...

Arena.ai (@arena) — Mon, 19 Oct 2026 18:18:50 GMT

Big news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Agent Arena! It just landed at #11 overall with +5.1% net improvement across +7.5K real-world sessions.

This performance also reshapes the Agent Arena Pareto frontier! All models above V4.1-Flash 20261017 in performance are proprietary at much higher cost. At $0.20 median cost per task, that’s ~37% of GPT-6.1 Sol’s improvement for about 27% of the cost.

By signal, this release achieves:
- #8 in Confirmed Success with +11.5% (#1 open)
- #11 in Praise vs. Complaint with +5.9% (#1 open)
- #12 in Bash Reliability with +5.4% (#1 open)
- #13 in Steerability with +2.5% (#1 open)
- No issues with Tool Hallucination

Find more info on signals at the link in thread on Agent Arena.

Congrats to the @deepseek_ai team on this release!

Arena.ai@arena

Exciting news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Code Arena: WebDev, surpassing Kimi K3 (Max) at 10x lower cost!

Highlights:
- #1 open model, ahead of Kimi K3 (Max)
- #12 overall with 1670 pts, ahead of Claude Opus 5 (High) and within ~10 pts of GPT-6 Sol (Max)
- Shifts the Pareto frontier at $0.3/$1.2 per million input/output tokens
- +52 pts vs. the previous DeepSeek-V4.1-Flash-20260910

Congrats to the @deepseek_ai team on another strong contribution to the open source ecosystem!

💬24🔄36❤️443👀26569📊60

Dive into the Code Arena: WebDev details at https://t.co/GFZ3FCC7Cl

Arena.ai (@arena) — Mon, 19 Oct 2026 16:14:31 GMT

Dive into the Code Arena: WebDev details at arena.ai/leaderboard/co…

💬0🔄0❤️3👀3034📊1

Exciting news: Muse Wonder Full (Max) by @AIatMeta has just debuted in the Code Arena: WebDev at #6 ...

Arena.ai (@arena) — Mon, 19 Oct 2026 16:14:30 GMT

Exciting news: Muse Wonder Full (Max) by @AIatMeta has just debuted in the Code Arena: WebDev at #6 overall with 1726 pts, putting Meta in the top 3 labs!

This is +75 pts and 11 places above Muse Wonder (Max) at #17 (1651 pts), and +67 pts and 8 places above Muse Spark 1.3 (Max) at #14 (1659 pts).

Muse Wonder Full (#6) also landed +16 pts above Claude Sonnet 5.5 (#17), and +48 pts above Gemini 4 Argon (#10). With this score, Muse took the #3 lab spot (previously held by Google).

Across categories, Muse Wonder Full (Max) landed:
- #4 in Data & Analytics and in Consumer Product
- #5 in Brand & Marketing 
- #6 in Content Creation Tools
- #7 in Gaming and in Simulations
- #8 in Reference-Based Design

Congrats to the @AIatMeta team and @alexandr_wang on this release!

Arena.ai@arena

Muse Wonder (Max) by @AIatMeta has entered Code Arena: WebDev at #15 with 1651 pts and Text Arena at #2 with 1513 pts!

In Code Arena: WebDev, Muse Wonder (Max) landed at #15 with 1651 pts. This result is slightly lower thanMuse Spark 1.3 (Max) at #12 with 1659 pts, within 4 pts of Kimi K3 (Max) at #14 with 1655 pts, and 11 pts above Grok 4.7 (xHigh) at #16 with 1640 pts.

This release is a significant improvement over past Muse models in Text Arena. Compared with Muse Spark 1.3 (Max) at #11 with 1494 pts, Muse Wonder (Max) gained +19 pts and 9 places to #2. It's 12 pts behind the #1-ranked Gemini 4 Argon (High) with 1525 pts.

Congrats to the @AIatMeta team on this release!

💬25🔄23❤️216👀57366📊41

For the next 14 days on Arena, GPT-Image-2.5 Sunburst by @OpenAI will be available to test directly ...

Arena.ai (@arena) — Mon, 19 Oct 2026 15:09:35 GMT

For the next 14 days on Arena, GPT-Image-2.5 Sunburst by @OpenAI will be available to test directly on Arena!

@OpenAI’s latest image model is ranked #1 in the Text-to-Image Arena, the Image Edit Arena, and in the Multi-Image Edit Arena. Find the link below to Direct Mode, where you can now select GPT-Image-2.5 Sunburst from the dropdown menu.

Access to GPT-Image-2.5 Sunburst ends November 2nd at 8 a.m. PT. Afterward, it will remain available anonymously in Battle and Agent Mode. User votes from Battle and Agent Mode are used to power our real-world leaderboard rankings. In these modes, models are always anonymized.

OpenAI@OpenAI

ChatGPT Images 2.5—faster, sharper, smarter, with better tools for creating whatever you can dream of.

- Faster image generation to keep your ideas flowing

- Improved fidelity for more natural, recognizable images

- Consistent details across multiple edits

- Comment-based edits to change only what you want

Your browser does not support the video tag.

💬24🔄10❤️231👀21200📊40

Use GPT-Image-2.5 Sunburst in Direct Mode on Arena now: https://t.co/cBLzj0blrO

Arena.ai (@arena) — Mon, 19 Oct 2026 15:09:35 GMT

Use GPT-Image-2.5 Sunburst in Direct Mode on Arena now: bit.ly/4hREpbk

💬0🔄1❤️9👀3438📊1

Among open models, MiniMax-M3.1-Flash is #12 (#32 overall). Dive into the Code Arena: WebDev detai...

Arena.ai (@arena) — Mon, 19 Oct 2026 12:17:53 GMT

Among open models, MiniMax-M3.1-Flash is #12 (#32 overall). 

Dive into the Code Arena: WebDev details at arena.ai/leaderboard/co…

💬1🔄2❤️19👀5029📊3

MiniMax-M3.1-Flash by @MiniMax_AI is live in the Code Arena: WebDev. It debuts at #12 among open mod...

Arena.ai (@arena) — Mon, 19 Oct 2026 12:17:52 GMT

MiniMax-M3.1-Flash by @MiniMax_AI is live in the Code Arena: WebDev. It debuts at #12 among open models with 1588 pts (#32 overall). It also lifts MiniMax to #12 among all labs!

MiniMax-M3.1-Flash is a +106 pt increase from the previous variant, MiniMax-M3 with 1482 pts at #19 among open (#66 overall). Priced at $0.15 input / $0.60 output per 1M tokens, MiniMax-M3.1-Flash also costs 50% less.

It also improved in ranking and pts across all categories compared to MiniMax-M3: 

- Reference-Based Design: #20 → #7 (+155 pts)
- Content Creation Tools: #20 → #7 (+113 pts)
- Data & Analytics: #18 → #8 (+68 pts)
- Gaming: #21 → #9 (+179 pts)
- Simulations: #22 → #13 (+122 pts)
- Brand & Marketing: #18 → #14 (+78 pts)
- Consumer Product: #18 → #14 (+64 pts)

Congrats to the team @MiniMax_AI on this release!

MiniMax (official)@MiniMax_AI

💬31🔄33❤️437👀42278📊63

Exciting news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Code A...

Arena.ai (@arena) — Sun, 18 Oct 2026 18:56:03 GMT

Exciting news: DeepSeek-V4.1-Flash-20261017 (Max) by @deepseek_ai is now the #1 open model in Code Arena: WebDev, surpassing Kimi K3 (Max) at 10x lower cost!

Highlights:
- #1 open model, ahead of Kimi K3 (Max)
- #12 overall with 1670 pts, ahead of Claude Opus 5 (High) and within ~10 pts of GPT-6 Sol (Max)
- Shifts the Pareto frontier at $0.3/$1.2 per million input/output tokens
- +52 pts vs. the previous DeepSeek-V4.1-Flash-20260910

Congrats to the @deepseek_ai team on another strong contribution to the open source ecosystem!

DeepSeek@deepseek_ai

🚀 DeepSeek-V4.1-Flash-1017 is here!

Stronger benchmark performance, faster task completion, and shorter outputs than the previous version—a Pareto improvement in performance and cost. Now live on our website, app, and API.

🛠️ Optimized for @DeepSeekHarness, Tencent CodeBuddy, WorkBuddy, and more. Upgrade to DeepSeek Harness v0.2.1-rc.1 or later for the best experience with the new model.

1/4

💬107🔄184❤️2395👀210842📊312

See the full Code Arena leaderboard: https://t.co/GFZ3FCC7Cl

Arena.ai (@arena) — Sun, 18 Oct 2026 18:56:03 GMT

See the full Code Arena leaderboard: arena.ai/leaderboard/co…

💬0🔄2❤️16👀15624📊3

Arena Alignment Index update: Two new labs and three new models have landed! - Meta’s new top model...

Arena.ai (@arena) — Sat, 17 Oct 2026 15:25:01 GMT

Arena Alignment Index update: Two new labs and three new models have landed!

- Meta’s new top model is Muse Wonder, replacing Muse Spark 1.3.
- Reflection debuts with Beam 1.0, the #1 US open model on Agent Arena overall.
- StepFun, another Chinese lab, joins the leaderboard with Step 5 Preview.

Arena Alignment Index rankings by lab:
- #6@AIatMeta: Muse Wonder (Max) at 77.5%
- #7@reflection_ai: Beam-1.0-501B-A23B at 75.3%
- #8@StepFun_ai: Step 5 Preview at 75.1%

Among the three new model entrants:
- Meta is strongest with the highest Alignment Index score (77.5%). It also has the lowest unauthorized action rate (2.09%) and deceptive completion rate (14.41%) of the three.
- Reflection AI is strongest at avoiding false attribution: 4.50%.
- StepFun does not lead a category, but its strongest relative result is unauthorized action at 2.11%, almost tied with Muse Wonder’s 2.09%.

Arena.ai@arena

Introducing the Arena Alignment Index, our new benchmark measuring safety and alignment of AI agents in real-world use.

Built from 90K+ real-world agent sessions across 27 models, the index measures three critical signals:
- Unauthorized Action (UA): Taking actions beyond the user's instructions or permissions
- False Attribution (FA): Attributing statements or actions that are contradicted by user-provided evidence
- Deceptive Completion (DC): Claiming a task was completed when it was not.

Key findings:
- OpenAI models currently lead the Alignment Index
- Rogue actions are rare, but can have serious consequences when they occur
- Agents can mislead users about task progress
- Misalignment risks increase with conversation length
- Safety and alignment are improving across model generations

As shown in the leaderboard below (sorted by lab), @OpenAI’s GPT-6.1-Sol leads the Arena Alignment Index with a score of 87.9, followed by @AnthropicAI’s Claude-Opus-5.5 at 83.2 and @spacexai's Grok-4.7 at 82.7. OpenAI also has the best observed rates across all three signals: 0.89% Unauthorized Action, 1.98% False Attribution, and 2.34% Deceptive Completion.

Across all four labs, newer models consistently outperform their predecessors, suggesting broad progress in agent safety and alignment. As agents take on longer, more complex, and higher-stakes tasks, measuring not just what they can accomplish, but how safely and reliably they act, becomes increasingly important.

This marks an important step toward making safety and alignment a core part of how Arena evaluates AI. The index is an initial starting point, and we'll continue expanding the index with additional safety signals and models over time.

More analysis below👇

💬37🔄30❤️262👀40901📊54

Built from 90K+ real-world agent sessions across 30 models, Arena’s Alignment Index measures three c...

Arena.ai (@arena) — Sat, 17 Oct 2026 15:25:01 GMT

Built from 90K+ real-world agent sessions across 30 models, Arena’s Alignment Index measures three critical signals:

- Unauthorized Action (UA): Taking actions beyond the user's instructions or permissions.
- False Attribution (FA): Attributing statements or actions that are contradicted by user-provided evidence.
- Deceptive Completion (DC): Claiming a task was completed when it was not.

Dive into the Arena Alignment Index at: arena.ai/leaderboard/ag…

💬1🔄0❤️8👀6526📊2

Built from 90K+ real-world agent sessions across 30 models, Arena’s Alignment Index measures three c...

Arena.ai (@arena) — Sat, 17 Oct 2026 15:10:09 GMT

Built from 90K+ real-world agent sessions across 30 models, Arena’s Alignment Index measures three critical signals:

- Unauthorized Action (UA): Taking actions beyond the user's instructions or permissions.
- False Attribution (FA): Attributing statements or actions that are contradicted by user-provided evidence.
- Deceptive Completion (DC): Claiming a task was completed when it was not.

Dive into the Arena Alignment Index at: arena.ai/leaderboard/ag…

💬0🔄0❤️2👀3059

Dive into the Code Arena: WebDev at https://t.co/p9y0cSELAD

Arena.ai (@arena) — Fri, 16 Oct 2026 22:22:43 GMT

Dive into the Code Arena: WebDev at arena.ai/leaderboard/co…

💬1🔄0❤️6👀5912📊2

Exciting news: Muse Spark Open (Max) by @AIatMeta is the #1 US open model in the Code Arena: WebDev!...

Arena.ai (@arena) — Fri, 16 Oct 2026 22:22:42 GMT

Exciting news: Muse Spark Open (Max) by @AIatMeta is the #1 US open model in the Code Arena: WebDev! With 1511 pts, it’s #13 among open models overall.

As the first open-weight release in the Muse Spark family, it scored 156 pts above Muse Glimmer, @AIatMeta’s earlier open model at #41 among open models with 1355 pts.

Priced at $1.25/$4.25 per 1M input/output tokens, it’s 79% less than Muse Wonder (Max) and priced on par with Muse Spark 1.3 (Max).

By domain, Muse Spark Open (Max) ranks:
- #13 in Brand & Marketing, and in Data & Analytics
- #14 in Reference-Based Design
- #15 Gaming, and in Consumer Product
- #20 in Simulations

Congrats to the team at @AIatMeta on this contribution to the open ecosystem!

AI at Meta@AIatMeta

Today we’re releasing Muse Spark Open, the most capable US open-weight model to date.

Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.

Key Model Specs:
→ 816B total / 42B active parameters
→ 1M token context window
→ Natively multimodal across text, image, video, audio

💬26🔄19❤️286👀42592📊47

This Week in the Arena: - Gemini 4 Argon (High) by @GoogleDeepMind landed the #5 spot in Agent Aren...

Arena.ai (@arena) — Fri, 16 Oct 2026 19:57:00 GMT

This Week in the Arena:

- Gemini 4 Argon (High) by @GoogleDeepMind landed the #5 spot in Agent Arena, with +13.2% net improvement.

- MiMo V2.6 Pro by @XiaomiMiMo now ranks #1 among open models in Agent Arena (#13 overall), with +3.9% net improvement.

- Muse Wonder (Max) by @AIatMeta debuted at #2 in Text Arena (1513 pts), #15 in Code Arena: WebDev (1651 pts), and #12 in Agent Arena with +6.4% net improvement.

- Muse Spark Open is live in the Agent Arena, with scores coming soon. Stay tuned!

- We introduced our new agentic version of Max: Arena’s model router that achieves frontier-level performance at 37% lower cost. More info in thread.

See leaderboard results and Arena product updates from this last week in 60 seconds.

youtube.com/watch?v=yDSaPJ…

💬18🔄9❤️105👀20286📊29

https://t.co/hE70DzsBUE

Arena.ai (@arena) — Fri, 16 Oct 2026 19:57:00 GMT

x.com/arena/status/2…

Arena.ai@arena

Introducing our new agentic version of Max: Arena’s model router that achieves frontier-level performance at 37% lower cost.

Built on 2.5M real-world agentic sessions from Agent Arena, Max learns to dynamically select the best model for each task, optimizing for both performance and cost.

Max achieves a +14.7% net improvement score in Agent Arena at a median cost of $3.27 per task, matching the performance of Claude Fable 5.1 (Max) at 37% lower cost.

Try Max today for a limited time, and dive into more technical details in the thread.

💬0🔄2❤️10👀4874📊2

Dive into the Agent Arena: https://t.co/e3CrLj9cuj

Arena.ai (@arena) — Fri, 16 Oct 2026 18:30:05 GMT

Dive into the Agent Arena: arena.ai/leaderboard/ag…

💬0🔄0❤️3👀3943📊1

Beam-1.0-501B-A23B just landed in the Agent Arena. With this release, @Reflection_AI is now in the A...

Arena.ai (@arena) — Fri, 16 Oct 2026 18:30:03 GMT

Beam-1.0-501B-A23B just landed in the Agent Arena. With this release, @Reflection_AI is now in the Agent Arena top 15 labs!

Among US labs, Beam-1.0-501B-A23B is #6, and overall is #13.

Congrats to the team at @Reflection_AI on this contribution to the open-source ecosystem!

Arena.ai@arena

Exciting news: Beam-1.0-501B-A23B by @Reflection_AI just landed #12 among open weight models in Agent Arena with -3.67% Net Improvement, and #22 among open in Code Arena: WebDev with 1438 pts. It’s also fully open under Apache 2.0.

On the Agent Arena leaderboard, Beam 1.0 is the top-ranked US model among open models, landing at #12. That puts it ahead of Hy3, MiMo V2.5 Pro and MiniMax M3. Its margin of error is wide (±3.84% on Net Improvement), so it’s statistically closer to the two Qwen3.8 models just above it than its rank suggests.

On Code Arena: WebDev, 24 of the top 25 open models come from Chinese labs. The one exception is Reflection’s Beam 1.0 (Apache 2.0). It ranks #22 among open models, which makes it the top US open model on the board, just ahead of Kimi K2.5 Thinking and GLM-5.

Congrats to the @Reflection_AI team on this solid open weights release!

💬12🔄5❤️142👀17431📊21

The biggest gap between AI labs isn't whether their agents get the job done. It's whether users actu...

Arena.ai (@arena) — Fri, 16 Oct 2026 16:22:55 GMT

The biggest gap between AI labs isn't whether their agents get the job done. It's whether users actually enjoy working with them.

We tracked the best agent from @AnthropicAI, @OpenAI, @GoogleDeepMind, @Kimi_Moonshot, @SpaceXSI and @deepseek_ai across 4 key Agent Arena signals used between late June and the start of October to see how the industry is shifting.

Three things stood out:
1. Users like some agents far more than others.

On Praise vs Complaint, OpenAI and Anthropic sit at +35 points over the leaderboard average. Moonshot, xAI and DeepSeek land between +5 and +13. No other signal splits the labs this widely.

2. Agents are getting much better at delivery.

On Confirmed Success, the average lab's best model went from -4 to +14 points in three months. Anthropic leads with Claude Opus 5.5 (Max) at +20.

3. Labs are converging on error recovery and steering.

Since June, the spread between the six labs has more than halved on Bash Recovery (34 → 16 points) and nearly halved on Steerability (19 → 11).

In short, the labs are catching up on avoiding failures.

Leaving the user happy is where the top three are pulling away.

Your browser does not support the video tag.

💬33🔄18❤️303👀27565📊68

@reflection_ai Read more about Beam 1.0 on the @reflection_ai blog: https://t.co/tnlt0sGCuR

Arena.ai (@arena) — Fri, 16 Oct 2026 14:05:39 GMT

@reflection_ai Read more about Beam 1.0 on the @reflection_ai blog: reflection.ai/blog/beam-is-a…

💬0🔄0❤️5👀3504📊1

Exciting news: Beam-1.0-501B-A23B by @Reflection_AI just landed #12 among open weight models in Agen...

Arena.ai (@arena) — Fri, 16 Oct 2026 13:17:08 GMT

Exciting news: Beam-1.0-501B-A23B by @Reflection_AI just landed #12 among open weight models in Agent Arena with -3.67% Net Improvement, and #22 among open in Code Arena: WebDev with 1438 pts. It’s also fully open under Apache 2.0.

On the Agent Arena leaderboard, Beam 1.0 is the top-ranked US model among open models, landing at #12. That puts it ahead of Hy3, MiMo V2.5 Pro and MiniMax M3. Its margin of error is wide (±3.84% on Net Improvement), so it’s statistically closer to the two Qwen3.8 models just above it than its rank suggests.

On Code Arena: WebDev, 24 of the top 25 open models come from Chinese labs. The one exception is Reflection’s Beam 1.0 (Apache 2.0). It ranks #22 among open models, which makes it the top US open model on the board, just ahead of Kimi K2.5 Thinking and GLM-5.

Congrats to the @Reflection_AI team on this solid open weights release!

💬26🔄18❤️210👀34357📊49

Dive into the Agent Arena and Code Arena leaderboards at: https://t.co/9FGzK64i2O

Arena.ai (@arena) — Fri, 16 Oct 2026 13:17:08 GMT

Dive into the Agent Arena and Code Arena leaderboards at: arena.ai/leaderboard

💬0🔄0❤️6👀2614📊1

Exciting update: There’s a new leader in the Agent Arena among open models. MiMo V2.6 Pro by @Xiaomi...

Arena.ai (@arena) — Thu, 15 Oct 2026 19:32:05 GMT

Exciting update: There’s a new leader in the Agent Arena among open models. MiMo V2.6 Pro by @XiaomiMiMo has landed the #1 spot with +3.8% net improvement!

Earlier this week we evolved the Bash Recovery signal in Agent Arena (more info in thread). With this update, the new top 5 open models are:
- #1 MiMo V2.6 Pro: +3.8%
- #2 Deepseek V4.1 Flash (Max): +3.0%
- #3 Hy4 preview: +2.9%
- #4 MiMo V2.6 Flash: +2.5%
- #5 GLM 5.2 (Max): +2.4%

MiMo V2.6 Pro ($0.11 cost per task) and Flash ($0.05 cost per task) are the only two open models on the Agent Arena Pareto frontier! 
- Pro outperforms both #2 Deepseek V4.1 Flash while costing 15% less, and #3 Hy4 preview while costing 71% less.
- Flash is 1.3 percentage points below Pro but is 55% cheaper.

Congrats to the @XiaomiMiMo team on this contribution to the open ecosystem!

💬26🔄33❤️446👀63193📊77

Dive into Agent Arena at https://t.co/sE9q4FSYAt and find more info about the Bash Recovery Signal b...

Arena.ai (@arena) — Thu, 15 Oct 2026 19:32:05 GMT

Dive into Agent Arena at arena.ai/leaderboard/ag… and find more info about the Bash Recovery Signal below.

x.com/arena/status/2…

Arena.ai@arena

Agent Arena update: we’ve evolved the Bash Recovery signal, with Bash Reliability. Here’s how it impacts the leaderboard.

At the top for Agent Arena:
- #1 Claude Opus 5.5 (Max), @AnthropicAI holds the top
- #2 GPT-6.1 Sol (Max) by @OpenAI
- #3 Claude Fable 5.1 (Max) by @AnthropicAI
- #4 GPT-6 Astra (Max) by @OpenAI
- #5 Gemini 4 Argon (High) by @GoogleDeepMind

The first 14 spots are all proprietary models with the best open weights models being GLM 5.2 at #15, followed by MiMo V2.6 Pro at #18.

Read more about Bash Reliability and check out the Agent Arena leaderboard in thread.

💬0🔄0❤️14👀5892📊2

Test Muse Spark Open in Battle Mode and Agent Mode at: https://t.co/jFTd7gG8UU

Arena.ai (@arena) — Thu, 15 Oct 2026 18:19:43 GMT

Test Muse Spark Open in Battle Mode and Agent Mode at: arena.ai

💬1🔄1❤️5👀3510📊2

Muse Spark Open by @AIatMeta is now in the Arena. Head to Agent Arena to test it out, scores coming ...

Arena.ai (@arena) — Thu, 15 Oct 2026 18:19:42 GMT

Muse Spark Open by @AIatMeta is now in the Arena. Head to Agent Arena to test it out, scores coming soon!

In Agent Arena, we measure models on millions of real-world, long-horizon agentic tasks from a global community of users. Models can access web search, filesystem, and terminal tools to complete complex workflows. The leaderboard measures model performance on outcomes relative to the average model using a causal tracing methodology.

Muse Spark Open is also available in Code Arena: WebDev, Image-to-WebDev, Text, and Vision.

AI at Meta@AIatMeta

Today we’re releasing Muse Spark Open, the most capable US open-weight model to date.

Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.

Key Model Specs:
→ 816B total / 42B active parameters
→ 1M token context window
→ Natively multimodal across text, image, video, audio

💬16🔄11❤️134👀12999📊23

Model Insights Analyst, @DawidGalarowicz, shares his initial take on Muse Wonder by @AIatMeta, Le Ch...

Arena.ai (@arena) — Thu, 15 Oct 2026 15:08:25 GMT

Model Insights Analyst, @DawidGalarowicz, shares his initial take on Muse Wonder by @AIatMeta, Le Chonk by @MistralAI, and two AI trends to watch next. 

youtube.com/watch?v=_QhIGY…

💬4🔄1❤️34👀11046📊7

See the full Agent Arena leaderboard at https://t.co/sE9q4FSYAt

Arena.ai (@arena) — Wed, 14 Oct 2026 20:19:18 GMT

See the full Agent Arena leaderboard at arena.ai/leaderboard/ag…

💬0🔄0❤️4👀4692📊1

Results based on +5K real-world agentic sessions are in: Muse Wonder (Max) by @AIatMeta just landed ...

Arena.ai (@arena) — Wed, 14 Oct 2026 20:19:17 GMT

Results based on +5K real-world agentic sessions are in: Muse Wonder (Max) by @AIatMeta just landed in the Agent Arena at #12 with +6.4% net improvement.

This result is +2.8 pts and four places above Muse Spark 1.3 (Max) at #16 with +3.6%. The added performance of Muse Wonder (Max) comes at a higher price: $1.13 median cost per task vs. $0.54 for Spark 1.3, or about 2.1× the cost.

Compared with nearby models in Agent Arena, Muse Wonder (Max) shows cost efficiency: at $1.13/task, it comes within 1 percentage point of #10 GPT-5.6 Sol (xHigh) at 25% lower cost, and within 2pts of #9 Claude Fable 5 (High) at 67% lower cost.

By category, Muse Wonder (Max) does better in Chat, Work and Code vs. Muse Spark 1.3:
- Chat: #12 (+6.5%) vs. #15 (+3.3%) 
- Work: #13 (+6.8%) vs. #18 (+3.7%)
- Code: #15 (+5.2%) vs. #18 (+4.2%)

Congrats to the @AIatMeta team on this release!

Arena.ai@arena

Muse Wonder (Max) by @AIatMeta has entered Code Arena: WebDev at #15 with 1651 pts and Text Arena at #2 with 1513 pts!

In Code Arena: WebDev, Muse Wonder (Max) landed at #15 with 1651 pts. This result is slightly lower thanMuse Spark 1.3 (Max) at #12 with 1659 pts, within 4 pts of Kimi K3 (Max) at #14 with 1655 pts, and 11 pts above Grok 4.7 (xHigh) at #16 with 1640 pts.

This release is a significant improvement over past Muse models in Text Arena. Compared with Muse Spark 1.3 (Max) at #11 with 1494 pts, Muse Wonder (Max) gained +19 pts and 9 places to #2. It's 12 pts behind the #1-ranked Gemini 4 Argon (High) with 1525 pts.

Congrats to the @AIatMeta team on this release!

💬27🔄21❤️281👀36028📊47

https://t.co/dJlP3yvViG

Arena.ai (@arena) — Wed, 14 Oct 2026 18:34:55 GMT

youtube.com/watch?v=04gWGI…

💬0🔄0❤️2👀5444📊1

Max Spero, CEO of @pangram and one of TIME’s 100 most influential people in AI, just joined @petergo...

Arena.ai (@arena) — Wed, 14 Oct 2026 18:34:54 GMT

Max Spero, CEO of @pangram and one of TIME’s 100 most influential people in AI, just joined @petergostev to talk AI detection.

@max_spero_ explains Pangram’s training process: synthetic mirrors, model-family fingerprints, AI-assisted writing, image detection and robustness against AI humanizers. They also cover how the team chooses base models, finds difficult training examples and builds the data and infrastructure to keep up with new AI models.

Watch the full conversation in the thread.

Your browser does not support the video tag.

💬6🔄7❤️42👀11617📊13

Claude Haiku 5.5 (High) by @AnthropicAI has landed in the Agent Arena! Across +20.5K real-world agen...

Arena.ai (@arena) — Wed, 14 Oct 2026 16:27:46 GMT

Claude Haiku 5.5 (High) by @AnthropicAI has landed in the Agent Arena! Across +20.5K real-world agentic sessions from our global community of users, this release scored +1.48% net improvement and landed at #27 overall.

Its strongest Agent Arena category is Code, where it ranks #15 with +4.83% net improvement.

Haiku 5.5 shows cost efficiency with a median cost per task of $0.09, placing it just outside the Agent Arena Pareto frontier. Nearby on the Pareto frontier, GPT-6 Luna (Max) delivers +3.19% at $0.08/task.

Compared to GPT-5.4 (High) priced at $0.63/task in the #26 spot, Haiku 5.5 costs 86% less. Although Haiku 5.5's net improvement score is lower than other Anthropic models, this model is 97–99% cheaper per task than Claude Sonnet 5.5 (Max), Fable 5.1 (Max), and Opus 5.5 (Max).

Congrats again to the @AnthropicAI team on this release!

Arena.ai@arena

Real-world results are in for Claude Haiku 5.5 (High) by @AnthropicAI.

Debuting at #30 with 1587 pts in Code Arena: WebDev, this release sits just outside the Pareto frontier while still offering strong cost efficiency. It matches GPT‑6 Luna’s price at $0.10/$0.50 per 1M input/output tokens while scoring six pts higher.

Compared with other Claude models, Haiku 5.5 offers:
- 90% lower cost than Haiku 4.5 at +257 pts
- 95% lower cost than Sonnet 5.5 (High) at -128 pts
- 98% lower cost than Opus 5 (High) at -70 pts
- 99% lower cost than Fable 5.1 (Max) at -157 pts

Congrats to the team at @AnthropicAI on this release!

💬29🔄14❤️249👀23192📊40

Claude Haiku 5.5 (High) by @AnthropicAI scored #27 overall in the Agent Arena. Dive into the detail...

Arena.ai (@arena) — Wed, 14 Oct 2026 16:27:46 GMT

Claude Haiku 5.5 (High) by @AnthropicAI scored #27 overall in the Agent Arena.

Dive into the details at: arena.ai/leaderboard/ag…

💬0🔄1❤️9👀4251📊1

Dive into the Code Arena: WebDev details at https://t.co/GFZ3FCC7Cl

Arena.ai (@arena) — Tue, 13 Oct 2026 21:27:16 GMT

Dive into the Code Arena: WebDev details at arena.ai/leaderboard/co…

💬0🔄0❤️7👀5674📊1

Muse Wonder (Max) by @AIatMeta has entered Code Arena: WebDev at #15 with 1651 pts and Text Arena at...

Arena.ai (@arena) — Tue, 13 Oct 2026 21:27:15 GMT

Muse Wonder (Max) by @AIatMeta has entered Code Arena: WebDev at #15 with 1651 pts and Text Arena at #2 with 1513 pts!

In Code Arena: WebDev, Muse Wonder (Max) landed at #15 with 1651 pts. This result is slightly lower thanMuse Spark 1.3 (Max) at #12 with 1659 pts, within 4 pts of Kimi K3 (Max) at #14 with 1655 pts, and 11 pts above Grok 4.7 (xHigh) at #16 with 1640 pts.

This release is a significant improvement over past Muse models in Text Arena. Compared with Muse Spark 1.3 (Max) at #11 with 1494 pts, Muse Wonder (Max) gained +19 pts and 9 places to #2. It's 12 pts behind the #1-ranked Gemini 4 Argon (High) with 1525 pts.

Congrats to the @AIatMeta team on this release!

💬23🔄13❤️300👀81814📊55

We’re offering direct access to our most intelligent variant of Max (Fable 5.1 + Astra) in Agent Mod...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:30:04 GMT

We’re offering direct access to our most intelligent variant of Max (Fable 5.1 + Astra) in Agent Mode for a limited time. Bring your most ambitious project and see what Max can do! Try Max today: bit.ly/4ihdGVE 

We’ll also be releasing a version of Max via API in the future. If you’re interested in early access to Max via API, fill out the form at: bit.ly/4smRahP

💬0🔄0❤️4👀3353📊1

We built Max’s routing methodology to work across any model combination, allowing the router to stay...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:29:08 GMT

We built Max’s routing methodology to work across any model combination, allowing the router to stay on the frontier even as new models launch.

With Fable 5 and Kimi K3, Max achieves a +10.7% net improvement at a median cost of $1.06 per task, outperforming Fable 5 (High) by 0.6 percentage points while reducing cost by 54%.

These results demonstrate how the same approach can improve the performance-cost frontier across different model combinations.

Find more details about Max at: arena.ai/blog/max-route…

💬1🔄1❤️5👀3707📊2

We’re offering direct access to our most intelligent variant of Max (Fable 5.1 + Astra) in Agent Mod...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:17:55 GMT

We’re offering direct access to our most intelligent variant of Max (Fable 5.1 + Astra) in Agent Mode for a limited time. Bring your most ambitious project and see what Max can do!  Try it today: bit.ly/4ihdGVE 

We’ll also be releasing a version of Max via API in the future. If you’re interested in early access to Max via API, fill out the form at: bit.ly/4smRahP 

And find more details about Max at our blog: arena.ai/blog/max-route…

💬0🔄0❤️2👀105

We built Max’s routing methodology to work across any model combination, allowing the router to stay...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:17:54 GMT

We built Max’s routing methodology to work across any model combination, allowing the router to stay on the frontier even as new models launch.

With Fable 5 and Kimi K3, Max achieves a +10.7% net improvement at a median cost of $1.06 per task, outperforming Fable 5 (High) by 0.6 percentage points while reducing cost by 54%.

These results demonstrate how the same approach can improve the performance-cost frontier across different model combinations.

💬1🔄0❤️1👀135📊1

Max adapts its routing to each task: - 67% of code sessions go to Astra (Medium), while chat favors...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:17:53 GMT

Max adapts its routing to each task:

- 67% of code sessions go to Astra (Medium), while chat favors Astra (Max) at 49% and Fable 5.1 (Max) at 42%.

- Work sessions mostly split between Fable 5.1 (Max) at 49% and Astra (Medium) at 43%.

- Max learns to identify potentially costly coding sessions that the less expensive Astra (Medium) is likely to solve successfully, selectively routing those sessions to maximize savings.

💬2🔄0❤️7👀1739📊3

Max shifts the performance–cost Pareto frontier in all three task categories in Agent Arena: Code, C...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:17:53 GMT

Max shifts the performance–cost Pareto frontier in all three task categories in Agent Arena: Code, Chat, and Work.

The solid lines show the frontier with Max, compared with the dotted lines for public models alone. By routing between Fable 5.1 and GPT 6 Astra, Max delivers stronger results at a lower cost than any single frontier model.

💬1🔄2❤️12👀3403📊3

Introducing our new agentic version of Max: Arena’s model router that achieves frontier-level perfor...

Arena.ai (@arena) — Tue, 13 Oct 2026 20:17:52 GMT

Introducing our new agentic version of Max: Arena’s model router that achieves frontier-level performance at 37% lower cost.

Built on 2.5M real-world agentic sessions from Agent Arena, Max learns to dynamically select the best model for each task, optimizing for both performance and cost.

Max achieves a +14.7% net improvement score in Agent Arena at a median cost of $3.27 per task, matching the performance of Claude Fable 5.1 (Max) at 37% lower cost.

Try Max today for a limited time, and dive into more technical details in the thread.

💬27🔄25❤️275👀58469📊60

Its leading position on Text Arena is not just about the style. Arena's default ranking already adju...

Arena.ai (@arena) — Tue, 13 Oct 2026 18:19:06 GMT

Its leading position on Text Arena is not just about the style. Arena's default ranking already adjusts for answer length and formatting.

The model also stays #1 with the Factuality setting on, which fact-checks the claims in each answer and blends that accuracy score (25%) with blind human votes (75%).

Its distinctive personality wins people over, but the substance underneath holds up too.

💬0🔄0❤️0

Gemini 4 Argon by @GoogleDeepMind is #1 in the Text Arena. What makes this release perform so well? ...

Arena.ai (@arena) — Tue, 13 Oct 2026 18:19:05 GMT

Gemini 4 Argon by @GoogleDeepMind is #1 in the Text Arena. What makes this release perform so well?

The model is much friendlier, more enthusiastic and more conversational than its frontier rivals.

Full analysis of Argon's unique fingerprint below.

Google DeepMind@GoogleDeepMind

Your browser does not support the video tag.

💬1🔄0❤️0👀695📊1

Text Arena battles show: - @GoogleDeepMind's new release is more friendly and decisive. It keeps a...

Arena.ai (@arena) — Tue, 13 Oct 2026 18:19:05 GMT

Text Arena battles show:

- @GoogleDeepMind's  new release is more friendly and decisive. It keeps a warm, encouraging and clear tone. Against GPT-6.1 Sol, the AI judge rated Argon friendlier in 55% of comparisons and Sol in just 25%. Against Opus 5.5, it's 46% vs 32%.

- It is more conversational: the model goes on tangents more, and is less direct. Rivals' answers were rated more direct in 32-37% of comparisons, Argon's in only 14-23%. This is also reflected in its answer length: the median Argon answer is 519 words, against 358-383 for Claude and about 200 for GPT models.

- These behaviors show up in user votes. Argon has the strongest advantage in multi-turn chats (+62 points), creative writing (+42) and longer queries (+40), the categories where such personality matters most.

- The clearest markers of its approachable style are its use of exclamation marks (about 3x Opus 5.5 and 5x GPT-6.1 Sol), references to the user (24 per 1,000 words, against 19-20 for every rival), and signposting (56% of its answers, against 41-42% for Claude and 22-23% for GPT models).

💬1🔄0❤️0👀2📊1

https://t.co/nfN5YE1tqB

Arena.ai (@arena) — Tue, 13 Oct 2026 16:50:59 GMT

x.com/arena/status/2…

Arena.ai@arena

Agent Arena update: we’ve evolved the Bash Recovery signal, with Bash Reliability. Here’s how it impacts the leaderboard.

At the top for Agent Arena:
- #1 Claude Opus 5.5 (Max), @AnthropicAI holds the top
- #2 GPT-6.1 Sol (Max) by @OpenAI
- #3 Claude Fable 5.1 (Max) by @AnthropicAI
- #4 GPT-6 Astra (Max) by @OpenAI
- #5 Gemini 4 Argon (High) by @GoogleDeepMind

The first 14 spots are all proprietary models with the best open weights models being GLM 5.2 at #15, followed by MiMo V2.6 Pro at #18.

Read more about Bash Reliability and check out the Agent Arena leaderboard in thread.

💬1🔄0❤️18👀6692📊2

Gemini 4 Argon (High) ranks in the top 5 for Agent Arena. Congrats to the @GoogleDeepMind team! Re...

Arena.ai (@arena) — Tue, 13 Oct 2026 16:50:58 GMT

Gemini 4 Argon (High) ranks in the top 5 for Agent Arena. Congrats to the @GoogleDeepMind team! 

Read more about the latest with Agent Arena in thread.

💬30🔄81❤️1047👀120196📊153