Which model tops arena.ai text arena Overall on Sep 21, 2026 12:00 ET?

Earlier updates

The chart shows each outcome’s published likelihood, including its baseline and news adjustment at that time.

1. Sep 15, 2026 UTC — Claude Fable 5 currently leads the Arena Elo rankings, while claude-fable-5.1-max sits in fifth place with a significant gap.; As of the latest data on September 15, the LMSYS Chatbot Arena leaderboard shows Claude Fable 5 in first place with an Elo score of 1,506, leading second-place Claude Opus 4.6 by just 3 points. The model in question, claude-fable-5.1-max, ranks only fifth with an Elo score of 1,498, trailing the leader by 8 points. With less than a week until the verification checkpoint, claude-fable-5.1-max would need to surpass four higher-ranked models to claim the top spot. Current evidence does not support it being in first place. The baseline likelihood is 77%, but direct news evidence shows it is not currently the leader, and while the gap is small, the path to overtaking is challenging. Based on this, the news evidence adjustment brings the final probability down from 77% to 67%, a decrease of 10.0 percentage points.
2. Sep 15, 2026 UTC — Claude Fable 5 firmly holds the top Arena Elo spot, while claude-fable-5.1-max remains in fifth with no sign of closing the gap.; Two independent leaderboard snapshots from this period both confirm that as of September 15, Claude Fable 5 holds the top spot on the LMSYS Chatbot Arena text leaderboard with an Elo of 1,506. claude-fable-5.1-max sits in fifth place on the Turing Score ranking with an Elo of 1,498, trailing the leader by 8 points and needing to overtake four higher-ranked models. The baseline likelihood is 81%, and after the news evidence adjustment, the final probability for claude-fable-5.1-max taking #1 is 71%, up 4.0 pp from the previous 67%. The direct news evidence shows no momentum for claude-fable-5.1-max. With less than a week until the checkpoint, it must leapfrog multiple strong competitors. Since the evidence is consistent with the prior period and offers no new support for a comeback, the -10 pp news evidence adjustment is maintained.
3. Sep 15, 2026 UTC — Claude Fable 5 continues to lead the Arena Elo rankings, while claude-fable-5.1-max remains in fifth place with no sign of catching up.; Two independent leaderboard snapshots this period reaffirm that as of September 15, Claude Fable 5 firmly holds the top spot on the LMSYS Chatbot Arena text leaderboard with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood stands at 81.5%, essentially unchanged from the prior period. Direct news evidence shows no momentum for claude-fable-5.1-max, which would need to overtake multiple strong competitors in under a week to claim the top spot. With evidence consistent with the prior period and no new developments supporting a reversal, the -10 percentage point news evidence adjustment is maintained. The final probability for claude-fable-5.1-max taking first drops from 81% last period to 71.5%, with Other model at 17.5%, claude-opus-5-max at 1.1%, claude-opus-5-high at 0.4%, and claude-opus-4-6-high at 0.3%.
4. Sep 15, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo rankings, while claude-fable-5.1-max stays in fifth with no sign of closing the gap.; Two independent leaderboard snapshots this period reconfirm that as of September 15, Claude Fable 5 holds the top spot on the LMSYS Chatbot Arena text leaderboard with an Elo score of 1,506. claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points and sitting behind four higher-ranked models. The baseline likelihood stands at 78%, down 3.5 percentage points from the prior period. Direct news evidence shows no momentum for claude-fable-5.1-max to close the gap; reaching first place would still require overtaking multiple strong competitors in under a week. With evidence consistent with the prior period and no new developments supporting a comeback, the -10 percentage point news evidence adjustment is maintained. In final, the probability of claude-fable-5.1-max taking first is 68%, down 3.5 percentage points from the prior 71.5%; Other model takes first is 18%, up 0.5 percentage points from 17.5%; claude-opus-5-max takes first is 0.7%, down 0.4 percentage points from 1.1%; claude-opus-5-high takes first is 0.4%, unchanged; and claude-opus-4-6-high takes first is 0.3%, unchanged.
5. Sep 15, 2026 UTC — Claude Fable 5 continues to lead the Arena Elo rankings, while claude-fable-5.1-max remains in fifth place with no sign of closing the gap.; Two independent leaderboard snapshots this period reaffirm that as of September 15, Claude Fable 5 sits firmly atop the LMSYS Chatbot Arena text rankings with an Elo of 1,506. claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points and facing four higher-ranked models ahead of it. The baseline likelihood stands at 75.5%, down 2.5 percentage points from the prior period. Direct news evidence shows no sign of claude-fable-5.1-max gaining ground; overtaking the top spot would require it to leapfrog multiple strong competitors in under a week. With evidence unchanged from the prior period and no new developments supporting a comeback, the -10 percentage point news evidence adjustment is maintained. This places the probability of claude-fable-5.1-max taking first at 65.5%, down 2.5 points from 68% last period; Other at 21.5%, up 3.5 points from 18%; claude-opus-5-max at 1.9%, up 1.2 points from 0.7%; claude-opus-5-high at 0.5%, up 0.1 points from 0.4%; and claude-opus-4-6-high at 0.3%, unchanged.
6. Sep 16, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots updated on September 15 reaffirm that Claude Fable 5 holds first place with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points and facing four higher-ranked models. The baseline likelihood sits at 80.5%. Direct news evidence shows no momentum for claude-fable-5.1-max, which would need to leapfrog multiple strong competitors in under a week to claim the top spot. With evidence unchanged from the prior period and no new developments supporting a reversal, the -10 percentage point news evidence adjustment is maintained. This yields a 70.5% probability for claude-fable-5.1-max to top the rankings, up 5.0 points from 65.5%; 20% for Other, down 1.5 points from 21.5%; 0.8% for claude-opus-5-max, down 1.1 points from 1.9%; 0.4% for claude-opus-5-high, down 0.1 points from 0.5%; and 0.3% for claude-opus-4-6-high, unchanged.
7. Sep 16, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 as the top-ranked model with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points and sitting behind four higher-ranked models. The baseline likelihood stands at 80.5%, and the direct news evidence shows no momentum for claude-fable-5.1-max, which would still need to leapfrog multiple strong competitors in less than a week. With evidence unchanged from the prior period and no new developments supporting a comeback, the -10 percentage point news evidence adjustment is maintained, resulting in a final probability of 70.5%, unchanged from the previous period. The probability of another model taking first place is 15.5%, down 4.5 pp from the previous 20%. The probabilities for claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high taking first place are 0.8%, 0.4%, and 0.3%, respectively, all unchanged.
8. Sep 16, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 in first place with an Elo score of 1,506. claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood sits at 84%, and the direct news evidence shows no momentum for claude-fable-5.1-max, which would still need to overtake multiple strong competitors in under a week. With evidence unchanged from the prior period and no new developments supporting a comeback, the -10 percentage point news evidence adjustment is maintained, leaving the final probability at 74%, up 3.5 percentage points from 70.5% last period. The probability of another model taking first place is 10.5%, down 5.0 percentage points from 15.5% last period. The probabilities for claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high are 1.1%, 0.4%, and 0.3%, respectively, with changes of +0.3, +0.0, and +0.0 percentage points.
9. Sep 16, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, again confirm Claude Fable 5 in first place with an Elo score of 1,506. claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max taking first is 89%, but direct news evidence shows it ranked only fifth, trailing the leader by 8 points with four stronger models ahead; the probability of overtaking all of them in under a week is far lower than the baseline likelihood reflects. The -10 percentage point news evidence adjustment is therefore maintained, yielding a final probability of 79%, up 5.0 percentage points from the prior period's 74%. The final probability for Other model takes first is 8.5%, down 2.0 percentage points from 10.5% in the prior period. claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high have final probabilities of 1.2%, 0.4%, and 0.3%, respectively.
10. Sep 16, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 as the top-ranked model with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to claim the top spot is 93%, but direct news evidence shows it sitting in fifth place, 8 points behind the leader with four stronger models in front—making a comeback in under a week far less probable than the baseline suggests. A -10 percentage point news evidence adjustment is maintained, yielding a final probability of 83%, up 4.0 percentage points from 79% in the prior period. The final probability for Other models stands at 6.5%, down 2.0 percentage points from 8.5%. Claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high sit at 1.2%, 0.4%, and 0.3% respectively, all unchanged from the prior period.
11. Sep 17, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 as the top-ranked model with an Elo score of 1,506. On the Turing Score leaderboard, claude-fable-5.1-max ranks fifth with an Elo of 1,498, trailing the leader by 8 points and sitting behind four higher-ranked models. The baseline likelihood for claude-fable-5.1-max to claim first place is 87%, but direct news evidence shows it is only fifth, 8 points behind, with four stronger models ahead. The probability of a comeback in under a week is far lower than the baseline reflects, so we maintain a -10 percentage point news evidence adjustment, yielding a final probability of 77%, down 6.0 points from the prior 83%. The final probability for Other models to top the ranking is 6.4%, down 0.1 points from 6.5%. Final probabilities for claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high remain unchanged at 1.2%, 0.4%, and 0.3%, respectively.
12. Sep 17, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 as the top-ranked model with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max taking first place is 86%, but direct news evidence shows it is only fifth and 8 points behind the leader, with four stronger models in front—making a comeback in under a week far less likely than the baseline suggests. We maintain a -10 percentage point news evidence adjustment, yielding a final probability of 76%, down 1.0 percentage point from the prior 77%. The final probability for Other model takes 1st is 10%, up 3.6 percentage points from the prior 6.4%. Final probabilities for claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high are 2.4%, 0.4%, and 0.3%, with changes of +1.1, +0.0, and +0.0 percentage points, respectively.
13. Sep 17, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 as the top-ranked model with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points and sitting behind four higher-ranked models. The baseline likelihood for claude-fable-5.1-max to take first place is 86%, but direct news evidence shows it is only fifth, 8 points behind, with four stronger models ahead. The probability of a comeback in under a week is far lower than the baseline suggests, so we maintain a -10 percentage point news evidence adjustment, yielding a final probability of 76%, unchanged from the prior judgment. The final probability for Other models to top the ranking is 6.7%, down 3.2 percentage points from 10% previously. Final probabilities for claude-opus-5-max, claude-opus-5-high, and claude-opus-4-6-high stand at 2.4%, 0.4%, and 0.3%, respectively, all unchanged.
14. Sep 18, 2026 UTC — Claude Fable 5 remains firmly atop the Arena Elo leaderboard, while claude-fable-5.1-max sits in fifth place with a widening gap, making a comeback highly unlikely.; Two independent leaderboard snapshots, both updated on September 15, reaffirm Claude Fable 5 in first place with an Elo score of 1,506. On the Turing Score leaderboard, claude-fable-5.1-max ranks fifth with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first place is 87%, but direct news evidence shows it is only in fifth place, 8 points behind the leader, with four stronger models ahead. The probability of a comeback in less than a week is far lower than the baseline reflects, so we maintain a -10 percentage point news evidence adjustment, yielding a final probability of 77%, up 1.0 percentage point from 76% in the prior period. The final probability for Other models to take first place is 5.5%, down 1.3 percentage points from 6.7% in the prior period. The final probability for claude-opus-5-max is 6.9%, up 4.6 percentage points from 2.4%. The final probabilities for claude-opus-5-high and claude-opus-4-6-high are 0.4% and 0.3%, respectively, both unchanged.
15. Sep 18, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period; the previous judgment is maintained. Two independent leaderboard snapshots from September 15 reaffirmed Claude Fable 5 in first place with an Elo of 1,506. On the Turing Score leaderboard, claude-fable-5.1-max ranked fifth with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first place is 87%, but direct news evidence shows it is only in fifth place, 8 points behind, with four stronger models ahead—making a comeback in under a week far less likely than the baseline suggests. The -10 percentage point news evidence adjustment is maintained, yielding a final probability of 77%, unchanged from the prior period. The final probability for Other models is 5.4%, down 0.1 pp from 5.5%. claude-opus-5-max sits at 3.4%, down 3.5 pp from 6.9%. claude-opus-5-high and claude-opus-4-6-high remain at 0.4% and 0.3%, respectively.
16. Sep 18, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period; the previous judgment is maintained. The two independent leaderboard snapshots from the prior period, both updated on September 15, reconfirm Claude Fable 5 in first place with an Elo of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first place is 89%, but the direct news evidence shows it sitting in fifth, 8 points behind the leader, with four stronger models to overtake in under a week—a scenario far less likely than the baseline suggests. The -10 percentage point news evidence adjustment is therefore maintained, yielding a final probability of 79%, up 2.0 points from the prior judgment's 77%. The final probability for Other to take first place is 8.6%, up 3.2 points from 5.4%. Claude-opus-5-max sits at 3.0%, down 0.4 points from 3.4%. Claude-opus-5-high and claude-opus-4-6-high remain at 0.4% and 0.3%, respectively, both essentially unchanged.
17. Sep 18, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period, so the previous judgment stands. The two independent leaderboard snapshots from last period, both updated on September 15, reaffirm that Claude Fable 5 holds first place with an Elo score of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first place is 87.5%, but the direct news evidence shows it is only in fifth place, 8 points behind the leader, with four stronger models to overtake. The probability of a comeback in less than a week is far lower than the baseline reflects, so we maintain a -10 percentage point news evidence adjustment, yielding a final probability of 77.5%, down 1.5 points from last period's 79%. The final probability for Other models to take first is 5.2%, down 3.3 points from 8.6%. Claude-opus-5-max's final probability is 3%, up 0.1 points from 3%. Claude-opus-5-high and claude-opus-4-6-high remain essentially flat at 0.4% and 0.3%, respectively.
18. Sep 18, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period, so the previous judgment is maintained. Two independent leaderboard snapshots from September 15 both confirm Claude Fable 5 in first place with an Elo of 1,506. The Turing Score leaderboard places claude-fable-5.1-max in fifth at 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to top the chart is 91.5%, but direct news evidence shows it is only fifth and 8 points behind, making a comeback in under a week far less likely than the baseline suggests. We maintain a -10 percentage point news evidence adjustment, yielding a final probability of 81.5%, up 4.0 points from the prior 77.5%. The final probability for Other models is 4%, down 1.2 points from 5.2%. claude-opus-5-max sits at 1.5%, down 1.6 points from 3%. claude-opus-5-high remains flat at 0.4%. claude-opus-4-6-high is at 0.1%, down 0.2 points from 0.3%.
19. Sep 19, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period, so the previous judgment is maintained. Two independent leaderboard snapshots from September 15 both confirm Claude Fable 5 in first place with an Elo of 1,506. claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first place is 94.5%, but the direct news evidence shows it in fifth place, 8 points behind the leader with four stronger models in front. The probability of a comeback in under two days is far lower than the baseline reflects, so the -10 percentage point news evidence adjustment is maintained, yielding a final probability of 84.5%, up 3.0 points from the previous 81.5%. The final probability for Other to take first place is 3.5%, down 0.5 points from 4%. The final probability for claude-opus-5-max is 2.8%, up 1.4 points from 1.5%. claude-opus-5-high holds at 0.4%, and claude-opus-4-6-high holds at 0.1%.
20. Sep 20, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period; the previous judgment is maintained. Two independent leaderboard snapshots from September 15, cited in the prior period, both confirm Claude Fable 5 in first place with an Elo of 1,506. The Turing Score leaderboard places claude-fable-5.1-max in fifth at 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first is 94.5%, but direct news evidence shows it sitting in fifth place, 8 points behind the leader, with four stronger models to overtake. The probability of a reversal in under two days is far lower than the baseline reflects, so the -10 percentage point news evidence adjustment is maintained, yielding a final probability of 84.5%, unchanged from the prior period. The final probability for Other model takes first is 3.5%, unchanged. claude-opus-5-max takes first sits at 3%, up 0.2 points from 2.8%. claude-opus-5-high takes first remains at 0.4%, and claude-opus-4-6-high takes first at 0.1%, both unchanged.
21. Sep 21, 2026 UTC — Claude Fable 5 remains firmly in first place, while claude-fable-5.1-max sits in fifth with a significant gap, making a comeback highly unlikely.; No new reports arrived this period; the previous judgment is maintained. The two independent leaderboard snapshots from the prior period, both updated on September 15, reconfirm Claude Fable 5 in first place with an Elo of 1,506. Claude-fable-5.1-max ranks fifth on the Turing Score leaderboard with an Elo of 1,498, trailing the leader by 8 points with four higher-ranked models ahead of it. The baseline likelihood for claude-fable-5.1-max to take first is 94.5%, but direct news evidence shows it is only in fifth place, 8 points behind, with four stronger models in front—making a comeback in under two days far less likely than the baseline suggests. The -10 percentage point news evidence adjustment is therefore maintained, yielding a final probability of 84.5%, unchanged from the prior period. The final probability for Other to take first is 2.1%, down 1.4 percentage points from 3.5%. Claude-opus-5-max's final probability is 2.7%, down 0.3 percentage points from 3%. Claude-opus-5-high's final probability is 0.2%, down 0.2 percentage points from 0.4%. Claude-opus-4-6-high's final probability is 0.1%, unchanged.
22. Oct 10, 2026 UTC — The final market result has been verified; this question is closed.; Final outcomes come from confirmed market winners or on-chain payouts, not a new model forecast.