[CRITICAL] Unknown plan class: GW-opencl-ati-3

Getting this in scheduler log when asking for O3AS work:

2025-01-22 21:29:45.4240 [PID=3416578]   Request: [USER#xxxxx] [HOST#13184820] [IP xxx.xxx.xxx.33] client 7.24.1
2025-01-22 21:29:45.4240 [PID=3416578] [debug]  [HOST#0] rpc_time:0 timezone:3600 d_total:268380930048.000000 d_free:189332979712.000000 d_boinc_used_total:2755349633.000000 d_boinc_used_project:2658643457.000000 d_boinc_max:0.000000
2025-01-22 21:29:45.4403 [PID=3416578] [debug]   have_master:0 have_working: 1 have_db: 1
2025-01-22 21:29:45.4403 [PID=3416578] [debug]   using working prefs
2025-01-22 21:29:45.4403 [PID=3416578] [debug]   have db 1; dbmod 1624117945.000000; global mod 1624117945.000000
2025-01-22 21:29:45.4403 [PID=3416578] [debug]   sending db prefs in reply
2025-01-22 21:29:45.4407 [PID=3416578]    [handle] [HOST#13184820] [RESULT#1717157725] [WU#866435239] got result (DB: server_state=4 outcome=0 client_state=0 validate_state=0 delete_state=0)
2025-01-22 21:29:45.4407 [PID=3416578]    [handle] cpu time 2823.719000 credit/sec 0.016930, claimed credit 47.805597
2025-01-22 21:29:45.4407 [PID=3416578]    [handle] [RESULT#1717157725] [WU#866435239]: setting outcome SUCCESS
2025-01-22 21:29:45.4430 [PID=3416578]    [send] effective_ncpus 14 max_jobs_on_host_cpu 999999 max_jobs_on_host 999999
2025-01-22 21:29:45.4430 [PID=3416578]    [send] effective_ngpus 1 max_jobs_on_host_gpu 999999
2025-01-22 21:29:45.4430 [PID=3416578]    [send] Not using matchmaker scheduling; Not using EDF sim
2025-01-22 21:29:45.4430 [PID=3416578]    [send] CPU: req 0.00 sec, 0.00 instances; est delay 0.00
2025-01-22 21:29:45.4430 [PID=3416578]    [send] ATI: req 51131.55 sec, 0.00 instances; est delay 0.00
2025-01-22 21:29:45.4430 [PID=3416578]    [send] work_req_seconds: 0.00 secs
2025-01-22 21:29:45.4430 [PID=3416578]    [send] available disk 97.43 GB, work_buf_min 84672
2025-01-22 21:29:45.4430 [PID=3416578]    [send] active_frac 0.999963 on_frac 0.458902 DCF 1.285427
2025-01-22 21:29:45.4439 [PID=3416578]    [mixed] sending non-locality work first (0.7395)
2025-01-22 21:29:45.4633 [PID=3416578]    [mixed] sending locality work second
2025-01-22 21:29:45.4648 [PID=3416578]    [send] send_old_work() no feasible result older than 336.0 hours
2025-01-22 21:29:45.8379 [PID=3416578]    [version] Checking plan class 'GW-opencl-ati-2'
2025-01-22 21:29:45.8402 [PID=3416578]    [version] reading plan classes from file '/BOINC/projects/EinsteinAtHome/plan_class_spec.xml'
2025-01-22 21:29:45.8403 [PID=3416578]    [version] WU#867246397 too new
2025-01-22 21:29:45.8403 [PID=3416578]    [version] Checking plan class 'GW-opencl-nvidia-2'
2025-01-22 21:29:45.8403 [PID=3416578]    [version] WU#867246397 too new
2025-01-22 21:29:45.8403 [PID=3416578]    [version] Checking plan class 'GW-opencl-ati-3'
2025-01-22 21:29:45.8403 [PID=3416578] [CRITICAL]   Unknown plan class: GW-opencl-ati-3
2025-01-22 21:29:45.8403 [PID=3416578]    [version] Checking plan class 'GW-opencl-nvidia-3'
2025-01-22 21:29:45.8403 [PID=3416578]    [version] parsed project prefs setting 'gpu_util_gw': 1.000000
2025-01-22 21:29:45.8403 [PID=3416578]    [version] No CUDA devices found
2025-01-22 21:29:45.8404 [PID=3416578]    [version] no app version available: APP#56 (einstein_O3AS) PLATFORM#9 (windows_x86_64) min_version 0
2025-01-22 21:29:45.8413 [PID=3416578] [debug]   [HOST#13184820] MSG(high) No work sent
2025-01-22 21:29:45.8413 [PID=3416578] [debug]   [HOST#13184820] MSG(high) see scheduler log messages on https://einsteinathome.org/host/13184820/log
2025-01-22 21:29:45.8413 [PID=3416578]    Sending reply to [HOST#13184820]: 0 results, delay req 60.00
2025-01-22 21:29:45.8413 [PID=3416578]    Scheduler ran 0.420 seconds

Any ideas why the server doesn't send me the new application version?