Gemini 3 Pro GPT-6 Astra Claude Fable 5.1 Head to head on 15 boards: GPT-6 Astra places best on 8 , Claude Fable 5.1 places best on 7 and Gemini 3 Pro places best on 1 (a tie counts for each). All 3 ranked on 8 of 19 boards.
Board Read Gemini 3 Pro GPT-6 Astra Claude Fable 5.1 Chat2 Text Arena Arena rating today today Nov 9 today today #18 1,486 #31 1,475 #6 1,504 best of the picked MultiChallenge Score today today Nov 9 today today #11 65.7% Reasoning7 HLE Accuracy today today Nov 9 today today #8 37.5% #1 54.8% best of the picked #2 46.5% Epoch ECI ECI today today Nov 9 today today #50 152.9 #1 169.1 best of the picked #5 166.9 ARC-AGI-2 Score today today Nov 9 today today #48 31.1% #1 95.0% best of the picked #7 90.0% LiveBench Global average today today Nov 9 today today #4 82.2% #1 83.4% best of the picked GPQA Diamond Accuracy today today Nov 9 today today #27 92.6% #2 95.8% best of the picked AA Intelligence Index today today Nov 9 today today #3 56.9 best of the picked #9 50.0 MMLU-Pro Accuracy · stale today today Nov 9 today today #2 90.1% Coding3 WebDev Arena Arena rating today today Nov 9 today today #68 1,440 #3 1,770 best of the picked #5 1,746 SWE-Bench Pro Resolved today today Nov 9 today today #4 90.2% #2 92.2% best of the picked SWE-bench Verified Resolved · stale today today Nov 9 today today #5 74.2% Agents and tools3 Search Arena Arena rating today today Nov 9 today today #18 1,200 #3 1,245 best of the picked #7 1,227 MCP Atlas Pass rate today today Nov 9 today today #24 70.3% #3 87.2% best of the picked BFCL Overall accuracy · stale today today Nov 9 today today #3 72.5% Vision2 Document Arena Arena rating today today Nov 9 today today #28 1,451 #11 1,477 #2 1,499 best of the picked Vision Arena Arena rating today today Nov 9 today today #12 1,289 best of the picked #17 1,285 #13 1,288 Speed1 AA Speed Output speed today today Nov 9 today today #16 21 tok/s #14 22 tok/s best of the picked Price1 AA Price Blended price today today Nov 9 today today #17 $20.0 best of the picked #17 $20.0 best of the picked
Green marks the best placing among the picked where two or more are ranked; ties are marked for each. Ranks are each board's own order (lower is better), scores as the board published them. Dash: not on that board.