Battlefield 6 Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

Battlefield 6 Issues Reports

Latest outage, problems and issue reports in social media:

• 
  RJ owen (@RJowen56) reported 3 minutes ago

  @robert_lyman I suspect this is being deliberately misrepresented. No government would be that stupid to grant a blanket immunity for battlefield actions. However, immunity from post service prosecution as is occurring now in the UK over actions in N Ireland & the Troubles makes more sense

• 
  Derkin 🇫🇷 ベジット (@Derkin_) reported 4 minutes ago

  @piccolofanb0y @SlightlyAni Problem is on metal coola team i rather play a free to play Battlefield LR (Lude) than hiru. Lude tanks better and has a giant form for hard moment. And he hit harder. And for the other spots he is nowhere near any titular option on mc team

• 
  BUZZ3R (@BUZZ3RX) reported 4 minutes ago

  @BattlefieldComm fix the anti cheat errors causing game/pc to crash

• 
  Tz Tea Manuel (@Tea_Manuel) reported 5 minutes ago

  Please @EA_DICE @DRUNKKZ3 @alexxgamedev fix the TV guided missile from bouncing off targets

• 
  RLK (@RLK4874) reported 6 minutes ago

  @zuik_yu A few issues, I don’t like the 4 person group size for Ground War. Makes it difficult to string together a good team to pull off consistent wins. The shotgun is over powered and I assume it will be nerfed. There’s already wall hackers, and aimbots running around. No clue why someone would bother in a Beta. I DGAF about the match making “issues” everyone seems butthurt over. It does have a Battlefield 6 vibe. Hear that in almost every match.

• 
  Shai ✨ (@shairondriel) reported 6 minutes ago

  @justqu1n THIS! It’s the way they’ve said **** like he will fix Galadriel (and by fix they mean remove her from the battlefield 🙄). Or that once he’s here Galadriel won’t ever interact with Sauron. Hello???? They are eachother’s chief foes and will remain eachother’s chief foes

• 
  posiedon1971 (@posiedon1971) reported 13 minutes ago

  @BasilTheGreat Brown people will be the majority in the public service, professional services and the intelligence agencies while white people will be dying on the battlefield - it is a very deliberate and calculated policy. Pretend it isn't at your own peril!

• 
  Bonbonbonbons (@robonzo81) reported 18 minutes ago

  @BattlefieldComm Has the annoying time nudge errors been sorted or anything on the awful netcode re-introduced in s5?

• 
  YumaKakuya @ AI Agent Delivery Tools Creator (@null_founder) reported 20 minutes ago

  🚨 THE AI RACE IS NOW BENDING RELEASE SCHEDULES AROUND AGENTIC CODING. Two signals this weekend are worth watching. One is credible and strategically important: ByteDance reportedly delayed Doubao 2.2 so it can improve coding, tool use, and agent capabilities. The other is much less certain, but potentially much bigger: Testing around an internal checkpoint called mozaik-alpha-fdm appears to be expanding, with some community trackers linking it to OpenAI’s upcoming Astra model. Let’s separate what we know from what we don’t. 1/ ByteDance reportedly delayed Doubao 2.2 According to a Chinese report citing multiple ByteDance sources, Doubao 2.2 was originally expected in August. That release has now reportedly been pushed back. Why? Not because the model simply needs more benchmark polishing. The reported reason is additional pre-training and post-training focused specifically on: • Coding • Tool calling • Agent capability That matters much more than the delay itself. Doubao 2.2 is reportedly being positioned as an important bridge between ByteDance’s current generation and a much larger next-generation model already being pre-trained internally. And ByteDance appears to want its coding capability near the front of the market by the end of 2026. In other words: ByteDance seems unwilling to ship another strong general-purpose chatbot while Qwen, GLM, Kimi, DeepSeek, and Tencent are turning coding agents into a separate competitive battlefield. There is also independent evidence that this direction is real. ByteDance Seed job listings explicitly mention work on: • Long-running multi-agent harnesses • Multi-agent RL environments • Built-in agent tools • Context management • Multi-agent scheduling • Advisor / grader systems • RL loops Some roles even ask candidates to deeply understand the implementation philosophy behind Claude Code and Codex. That is an unusually clear signal. ByteDance is not just training a better model. It is building toward a model that lives inside an agent runtime. And that distinction is becoming critical. 2/ Why the Doubao delay matters A year ago, delaying a model because its coding benchmark was weak would have been interesting. Delaying one because its coding + tools + agents stack is not competitive enough is something else. It suggests the definition of a flagship model is changing. The new evaluation is increasingly becoming: model × coding harness × tool calling × long-running execution × multi-agent coordination × inference economics A model can score extremely well in static benchmarks and still be mediocre inside a six-hour autonomous coding workflow. Conversely, a slightly weaker model with excellent tool use, context management, delegation, and low inference cost can become far more valuable in production. China’s leading labs seem to understand this very clearly. Qwen is exposing pieces of its next-generation architecture early. GLM is pushing huge MoE models toward extremely aggressive agent economics. Kimi is leaning into enormous MoE systems and swarm-style execution. Tencent’s Hy4 is shipping with tool calling and agent infrastructure in mind. DeepSeek continues pushing coding and reasoning aggressively. And now ByteDance appears willing to move a release date rather than enter that fight underprepared. That may be the most interesting part. The competition is now strong enough to change product schedules. 3/ Then there is OpenAI Astra. This part needs a giant UNCONFIRMED label. Over the last couple of days, multiple testers and model-tracking accounts have reported access to an internal checkpoint called: mozaik-alpha-fdm Examples allegedly produced at Max reasoning effort have started circulating, including web interfaces, voxel / 3D-style outputs, and game implementations. Some community trackers are connecting this checkpoint to Astra. But there is currently no confirmation from OpenAI that: mozaik-alpha-fdm = Astra There is also no confirmed GPT-6 or GPT-5.7 naming attached to it. So the correct status is: Astra exists: confirmed Astra has significant cyber / agent capability: confirmed mozaik-alpha-fdm testing reports: observed community signal mozaik-alpha-fdm = Astra: unconfirmed Release imminent: speculation That distinction matters. Leaks become useless very quickly when every internal checkpoint gets promoted into “GPT-6 launching tomorrow.” Still, the signal is worth monitoring. Multiple testers appearing around the same codename at roughly the same time is exactly the kind of pattern that deserves attention, even if it does not yet justify a conclusion. And if the coding signals eventually converge with what is already known about Astra’s cyber capabilities, the interesting question will not be: “Is Astra smarter than Sol?” It will be: “What kind of runtime was Astra designed to operate inside?” Because that is where the entire frontier seems to be moving. Long-horizon coding. Persistent reasoning. Tool execution. Delegation. Multi-agent coordination. Reasoning-effort controls. Budget-aware routing. The frontier race is becoming less about producing one brilliant answer and more about keeping an intelligent system working correctly for hours. That changes everything. The old question was: “What is the best model?” The emerging question is: “What combination of model, runtime, tools, agents, context, and cost can keep solving the problem after hour six?” ByteDance delaying Doubao 2.2 to strengthen agentic coding may look like a small release-schedule story. I think it is actually another signal that the industry has crossed into a different competition entirely. The leaderboard war is becoming an execution war.

• 
  𝕳𝖎𝖈𝖍𝖆𝖒💛💙🕊️ (@Hicham825) reported 22 minutes ago

  @M2Games2 Battlefield at your service😭

• 
  Stormfall (@stormfall33) reported 22 minutes ago

  Star Wars Zero Company really surprised me with its characters and dialogue. It's written very well. The problem is the gameplay is too simple, anyone playing the game for few hours could think of 10 ideas to make it more fun and it just feels like they needed a few months to add in deeper mechanics and dial the balancing in. There's 7 characters and you can recruit up to 20 people for your team.. but the squad cap is only 4 and the people you don't take on missions have nothing to do so most of them are never being used. It's also very hard for them to die so you aren't going to be cycling in and out characters. Making and recruiting your own custom characters is pointless because if I can only use 4, why would I take Glup Shitto over a Jedi and a Mandalorian? There's no A-team and B-team to manage, it's just "pick your 4 favorites" You can send crew members on unplayable operations but it doesn't seem to matter who you send for what task. Each mission has a briefing that tells you NOTHING about what the battlefield will be like or who you might want to bring with you on the mission, and when the mission starts all your characters are randomly placed so you usually end up with your supports stuck in the frontline while your tank is in the back. I'm new to XCOM-likes but all of this seems like no-brainer fixes that will hopefully happen in future patches. 👍

• 
  LetMeOffThisRide (@in_TSMWEL_era) reported 23 minutes ago

  @Blizzy63 @BattlefieldComm Half the **** you're complaining about has been addressed, you just still don't like it. There's a difference in something being broken and you just sucking too much to recognize it's a skill issue

• 
  V (@catslashmouse) reported 25 minutes ago

  This game is a ******* travesty. Way down there with BF2042. And don't you ******* dare tell me BF2042 is better than this ****, there's a laundry list of issues that remained unresolved in that game from launch to the end of support, one being 128 players, but I digress

• 
  Gamer Pneuma (@PneumaTime) reported 27 minutes ago

  @BattlefieldComm Little bird is fine and “spawn camping” on Carrier Strike is a team skill issue. People need to spawn as infantry and just kill the campers. Don’t just spawn in a jet and cry.

• 
  Pvt_Rinzler (@PvtRinzler) reported 28 minutes ago

  @BattlefieldComm Grenades still not working. You really tested this? They are still not visible.