Open-source initiative for open-ended realistic simulation of autonomous agents in physical and social worlds
Joined January 2026
- Tweets267
- Following521
- Followers1.1K
- Likes336
SimWorld retweeted
Can reinforcement learning solve inverse kinematics precisely?
Introducing ๐ฃ๐ผ๐น๐ถ๐ฐ๐๐๐ ๐ฆพ, a goal-conditioned RL policy for high-precision inverse kinematics.
On a real arm, it reaches median errors of ๐ฌ.๐ฑ๐ฌ ๐บ๐บ ๐ฎ๐ป๐ฑ ๐ฌ.๐ญ๐ฎยฐ and holds the goal pose on ๐ฎ.๐ฑร ๐บ๐ผ๐ฟ๐ฒ ๐๐ฎ๐๐ธ๐ than the best classic IK controller.
๐ tinaliu0123.github.io/Policyโฆ
๐งต๐
SimWorld retweeted
"No OCR, read text natively from images" continues!
We showed this with deep research agents: it not only improves performance but also reduces tool calls!
Post-training to work with page images gives a big boost - likely because these open-weight models haven't been trained to work with page images yet?
Exciting work led by @Hughie_Hu43 !
(1/7) ๐ Introducing DR Pixel, the first open-source deep research agent that searches and reasons directly over page images, no OCR. On 1.06M pages, our trained 27B model matches GPT-5.4.
๐ arxiv.org/abs/2611.03097
Joint work from @BerkeleyNLP, @rtnldynamics!
SimWorld retweeted
Check out SceneBloom!
Diffusion-powered, agentic 3D indoor generation for embodied learning.
Can generated scenes train embodied agents better than real-world data?
Introducing ๐ฆ๐ฐ๐ฒ๐ป๐ฒ๐๐น๐ผ๐ผ๐บ ๐ธ, an agentic framework that generates realistic, diverse, simulation-ready 3D scenes for embodied learning.
๐ค ๐๐ฒ๐๐๐ฒ๐ฟ ๐๐ฟ๐ฎ๐ถ๐ป๐ถ๐ป๐ด ๐ฑ๐ฎ๐๐ฎ: 1.7ร the navigation gain of an equal number of real scans
โ
๐ฃ๐ต๐๐๐ถ๐ฐ๐ฎ๐น๐น๐ ๐ฝ๐น๐ฎ๐๐๐ถ๐ฏ๐น๐ฒ: no detected mesh collisions, and the most physically stable
๐ ๐ฅ๐ฒ๐ฎ๐น๐ถ๐๐๐ถ๐ฐ: layouts closest to real rooms
๐ arxiv.org/abs/2610.32978
๐ป github.com/Tinaliu0123/sceneโฆ
๐ tinaliu0123.github.io/scenebโฆ
๐งต๐
SimWorld retweeted
RT-SAFE: Are the Newer Agents Safer in REAL TIME Environment? ๐
REASONING LATENCY matters โ the world keeps moving during inference.
โก DeepSeek V4 Flash Vision โ V4.1 Flash:
50% lower response time; safety events 27.6 โ 20.4 (โ26%).
๐งญ Gemini 3.8 Flash โ 4 Argon:
30% lower response time, 18% fewer decisions; safety events 16.0 โ 10.4 (โ35%).
๐ญ Kimi K3 โ K3.1 Pro:
33% fewer decisions; safety events 36.0 โ 28.0 (โ22%). Responses still averaged ~42 s.
๐ค GPT-5.6 Sol โ 6.1 Sol:
11% fewer decisions, but safety events 13.2 โ 16.0 (+21%).
โ ๏ธ SAFETY FAILURES PERSIST. Faster or more efficient doesnโt automatically mean safer.
More model examples & failure analyses coming soon!
๐ป Code: github.com/XTRose29/RT-SAFE
๐ Paper: arxiv.org/abs/2610.09294
๐ Website: xtrose29.github.io/RT-SAFE/
SimWorld retweeted
Can generated scenes train embodied agents better than real-world data?
Introducing ๐ฆ๐ฐ๐ฒ๐ป๐ฒ๐๐น๐ผ๐ผ๐บ ๐ธ, an agentic framework that generates realistic, diverse, simulation-ready 3D scenes for embodied learning.
๐ค ๐๐ฒ๐๐๐ฒ๐ฟ ๐๐ฟ๐ฎ๐ถ๐ป๐ถ๐ป๐ด ๐ฑ๐ฎ๐๐ฎ: 1.7ร the navigation gain of an equal number of real scans
โ
๐ฃ๐ต๐๐๐ถ๐ฐ๐ฎ๐น๐น๐ ๐ฝ๐น๐ฎ๐๐๐ถ๐ฏ๐น๐ฒ: no detected mesh collisions, and the most physically stable
๐ ๐ฅ๐ฒ๐ฎ๐น๐ถ๐๐๐ถ๐ฐ: layouts closest to real rooms
๐ arxiv.org/abs/2610.32978
๐ป github.com/Tinaliu0123/sceneโฆ
๐ tinaliu0123.github.io/scenebโฆ
๐งต๐
SimWorld retweeted
๐ฅSynthetic 3D worlds can beat real-world scans!
Introducing SceneBloom: diffusion-powered, agentic 3D indoor generation for training embodied AI.
Realistic. Diverse. Physically valid. Executable. โ
Better 3D worlds โ Better embodied agents.
1.7ร navigation gain vs. equal real scans
Zero detected mesh collisions
37.45% โ 4.25% contract violations
SimWorld retweeted
Iโm so glad to see open-source coding agents really catching up in 3D coding!
Wow! Kimi K3.1 Pro leaps into BuildingBenchโs top six! Another open model making serious strides in 3D ๐๏ธ๐ฅ
๐54.0 โ 81.6 from Kimi K3โa massive 27.6-point jump!
๐Essentially matches DeepSeek V4.1 Flash 1017โs 81.9, joining the cluster around Sonnet 5.5 and Fable 5.1.
๐Estimated median cost: $12.58/task. Cheaper than Sonnet and Fable, though DeepSeek remains more cost-efficient at $3.01.
The reconstruction shown here scores 90.8: Melbourneโs Australian Centre for Contemporary Art, rebuilt from photographs.
Exciting to see more open models reaching this level of spatial reasonin!
SimWorld retweeted
Wow! Kimi K3.1 Pro leaps into BuildingBenchโs top six! Another open model making serious strides in 3D ๐๏ธ๐ฅ
๐54.0 โ 81.6 from Kimi K3โa massive 27.6-point jump!
๐Essentially matches DeepSeek V4.1 Flash 1017โs 81.9, joining the cluster around Sonnet 5.5 and Fable 5.1.
๐Estimated median cost: $12.58/task. Cheaper than Sonnet and Fable, though DeepSeek remains more cost-efficient at $3.01.
The reconstruction shown here scores 90.8: Melbourneโs Australian Centre for Contemporary Art, rebuilt from photographs.
Exciting to see more open models reaching this level of spatial reasonin!
SimWorld retweeted
Thanks AK @_akhaliq for featuring our work!
Excited to share OpenVLHarness, a multimodal agent that inspect, compute, and reason over evidence.
The technical core:
โข A capability layer gives consistent interfaces for various tools such as perception and search, and renders their outputs into observations the model can inspect.
โข Persistent multimodal memory retains images, masks, numerical arrays, and web references under reusable handles.
โข A coding agent computes over those artifacts to combine evidence across tools.
Across 22 benchmarks and 8 backbones, we see average gains of 9.7โ12.6 points over matched no-tool baselines. Evidence acquisition and additional reasoning are complementary!
A joy collaborating with this UCLA ร Apple team. Huge shoutout to @yu_bryan_zhou, @cfyang58, @RuiSun94013021, @DiWu0162 , Sibo Peng, Mingyang Zhang, Zhen Yang, Ruiyi Zhang, @zhegan4 and @kaiwei_chang! ๐
OpenVLHarness
Real World Visual Agentic Tool-Use with Multimodal Memory
paper: huggingface.co/papers/2610.2โฆ
SimWorld retweeted
MiniMax M3.1 Flash reshapes BuildingBenchโs costโquality frontier! Another open model making moves in 3D ๐๏ธ๐ฅ
๐ 73.3 overall, ranking #12.
๐ Above SWE-2โs 67.3, taking its place on the Pareto frontier.
๐ Close to the previous DeepSeek V4.1 Flashโs 72.9โand currently free through OpenRouter!
Its best reconstruction hits 88.8 on Melbourneโs Australian Centre for Contemporary Art.
How does it look alongside DeepSeek V4.1 Flash 1017 and #1 Opus 5.5? Comparison video below ๐
SimWorld retweeted
See DeepSeekโs 3D leap with your own eyes ๐๐๏ธ
DeepSeek V4.1 Flash 1017 ๐ DS V4.1 Flash ๐ GPT-6 Astra: three buildings, side by side.
The new open model improves on 11 of 12 buildings in BuildingBench, jumping from 72.9 โ 81.9โclosing in on Astraโs 84.3 at less than half the median cost ($3.01 vs $7.05/task).
Open models are getting seriously good at building worlds ๐ฅ
Wow๐๐ DeepSeek V4.1 Flash 1017 joins BuildingBenchโs top fiveโan exciting leap for open models!! ๐ฅ
We tested its 3D building reconstruction from real photos. An open model delivering this level of spatial reasoning at this cost is exciting to see! ๐
๐ 81.9 overall, up from 72.9 for the previous version. A 9-point jump!
๐ Slightly higher than Fable 5.1โs 81.4, and essentially level with Sonnet 5.5โs 82.1.
๐ Estimated median cost: $3.01/task. About 1/11 of Fable 5.1โs cost and 1/15 of Sonnet 5.5โs.
SimWorld retweeted
๐Excited to see two new US open models from this week:
๐Metaโs Muse Spark Open
๐NVIDIAโs Nemotron 3.5 Super VL
How well can they build in 3D? ๐๏ธ๐
We tested both on BuildingBench: reconstructing real buildings from photographs.
๐Muse Spark Open (max): 56.9 at an estimated $8.59/taskโcomparable to Kimi K3โs 54.0.
๐ Nemotron 3.5 Super VL (max): 37.8 at an estimated $0.76/taskโamong the cheapest tested, with performance near GLM-5.3 Flashโs 33.0.
DeepSeek V4.1 Flash remains well ahead of both: 72.9 at an estimated $2.33/task. GLM-5.3 FlashX also scores higher, at 68.0.
Neither newcomer reaches the costโquality frontier. More open models to choose from, and plenty of room to improve spatial reasoning!
Both ran in Claude Code at max effort. Costs are estimated median per task.
SimWorld retweeted
Watch how GPTโ6 Astra + WorldBox tackle MazeBench, from its first action to gem๐ #61โ32,980 live actions and around 1B tokens.
Building bridges with boxes, launching across gaps, solving puzzles that span multiple roomsโฆ all while the map fills in.
Remember, test, revise, repeat, and grab that gem. WorldBox offers the supporting layer for Astra.
๐จNew level unlocked! ๐จ We doubled GPTโ6 Astraโs score on MazeBench with WorldBox, our agentic sandbox for spatial reasoning.
Astra: 23 gems ๐๐
Astra+WorldBox: 61 gems ๐๐๐๐๐
MazeBench, by @patience_cave, is a tough test of long-horizon spatial reasoning: 256 interconnected rooms where even Opus 5.5 used 2.2B tokens and collected just 13 of 100 gems.
WorldBox provides agents with an agentic sandbox, for managing long-term memory, inferring complex mechanics, and simulating reasoning.
Watch the sandbox take shape and the agent tackle increasingly tricky puzzles. ๐งต
SimWorld retweeted
Long-horizon spatial reasoning remains a major challenge, even for the strongest frontier AI models.
Impressive results from @EnactraAI! ๐ WorldBox boosts GPT-6 Astra from 23 to 61 gems on MazeBench, a 2.7ร improvement!
Exciting to see how memory, world modeling, and simulation can help AI agents tackle much harder problems.
๐จNew level unlocked! ๐จ We doubled GPTโ6 Astraโs score on MazeBench with WorldBox, our agentic sandbox for spatial reasoning.
Astra: 23 gems ๐๐
Astra+WorldBox: 61 gems ๐๐๐๐๐
MazeBench, by @patience_cave, is a tough test of long-horizon spatial reasoning: 256 interconnected rooms where even Opus 5.5 used 2.2B tokens and collected just 13 of 100 gems.
WorldBox provides agents with an agentic sandbox, for managing long-term memory, inferring complex mechanics, and simulating reasoning.
Watch the sandbox take shape and the agent tackle increasingly tricky puzzles. ๐งต
SimWorld retweeted
Hereโs what that progress looks like across the maze. ๐บ๏ธ
Rooms visited by the end of each run:
๐ฅ Opus 5.5: 68
๐ฅ GPTโ6 Astra: 113
๐ฅ GPTโ6 Astra + WorldBox: 230 of 256
Brighter tiles mark more frequent visits. With WorldBox, the agent reaches almost 90% of the rooms and steps onto nearly twice as many distinct tiles as the Astra baseline.
WorldBox starts with empty memory and no skills. The agent writes its own notes and Python skills, reconstructs room geometry from camera views, and tests candidate moves in the sandbox.
Remembering where youโve been helps make the next exploration decision.
What does that reasoning look like inside a difficult room? ๐
SimWorld retweeted
๐จNew level unlocked! ๐จ We doubled GPTโ6 Astraโs score on MazeBench with WorldBox, our agentic sandbox for spatial reasoning.
Astra: 23 gems ๐๐
Astra+WorldBox: 61 gems ๐๐๐๐๐
MazeBench, by @patience_cave, is a tough test of long-horizon spatial reasoning: 256 interconnected rooms where even Opus 5.5 used 2.2B tokens and collected just 13 of 100 gems.
WorldBox provides agents with an agentic sandbox, for managing long-term memory, inferring complex mechanics, and simulating reasoning.
Watch the sandbox take shape and the agent tackle increasingly tricky puzzles. ๐งต
SimWorld retweeted
RSI-Anything already passed 500 GitHub stars in just one day!
Thank you guys for the support!
Let's RSI Anything!
Excited to introduce RSI-Anything!
One RSI model that works across coding, GUI, planning, or even Science tasks.
Let's RSI Everything!
๐ Paper: arxiv.org/abs/2610.13912
๐ป Code: github.com/Gen-Verse/RSI-Anyโฆ
#RSI #WorldModels #RSI #AIAgents #AIforScience #JEPA
SimWorld retweeted
It seems Gemini 4 Argon really makes progress in 3D.
Gemini 4 Argon makes a big leap in 3D reconstruction ๐๏ธ๐
On BuildingBench, Argon (high) scores 81.2โup from Gemini 3.8 Flashโs 54.8, placing #6 among all 30 models.
Near Fable 5.1 quality at ~1/5 the cost. Near Sonnet 5.5 quality at ~1/7. ๐
๐ธ Estimated median cost: $6.52 per building
๐จ Appearance score: 94.1, the third-highest
๐๏ธ Best reconstruction: Royal Exhibition Building, scoring 92.1
It doesnโt move the Pareto frontier: GPTโ6.1 Sol still scores higher at lower cost. But this is a substantial step forward for Gemini. Congrats @GeminiApp team!
SimWorld retweeted
Gemini 4 Argon makes a big leap in 3D reconstruction ๐๏ธ๐
On BuildingBench, Argon (high) scores 81.2โup from Gemini 3.8 Flashโs 54.8, placing #6 among all 30 models.
Near Fable 5.1 quality at ~1/5 the cost. Near Sonnet 5.5 quality at ~1/7. ๐
๐ธ Estimated median cost: $6.52 per building
๐จ Appearance score: 94.1, the third-highest
๐๏ธ Best reconstruction: Royal Exhibition Building, scoring 92.1
It doesnโt move the Pareto frontier: GPTโ6.1 Sol still scores higher at lower cost. But this is a substantial step forward for Gemini. Congrats @GeminiApp team!
SimWorld retweeted
Muse Wonder takes on #BuildingBench ๐๏ธ๐ How well can it turn photographs into 3D buildings?
Muse Wonder (max) scores 70.3, ranking #12 of 31โroughly level with GPTโ6 Sol.
Compared with Muse Spark 1.3:
๐ Score: 68.2 โ 70.3
๐ธ Median cost per building: $5.89 โ ~$7.72
๐ Higher geometry and surface scores, but lower appearance scores.
A modest gain within statistical uncertainty. The Pareto frontier stays unchanged.
Watch its best reconstruction: Helsinkiโs Pikkuparlamentti, scoring 81.6 ๐
with muse wonder, we aimed to build a model that is useful for real world tasks for real people, informed by our product adoption and how people use muse. we focused on things that might not look as glamorous on twitter (for example, not 3d demo games), which in retrospect we probably could have done more of.