Rapid releases from OpenAI and Anthropic, including GPT-5.6 Sol's July 2026 launch and Claude Fable 5's strong agentic coding results, have driven frontier large language models to 95%+ on SWE-Bench Verified and leading Elo ratings in live Coding Arenas. Chinese labs have accelerated competition, with Kimi K3 topping frontend code arenas and DeepSeek V4, GLM-5.2 matching or exceeding proprietary models on agentic benchmarks like Terminal-Bench. This momentum, fueled by iterative model scaling, tool-use improvements, and open-weight gains, supports trader consensus that current leaders or near-term successors will clear elevated arena thresholds by year-end, though benchmark saturation and potential timeline slips remain key uncertainties for resolution.
Résumé expérimental généré par IA à partir des données Polymarket. Ceci n'est pas un conseil de trading et ne joue aucun rôle dans la résolution de ce marché. · Mis à jour$185,152 Vol.
1560
44%
1580
24%
1600
14%
$185,152 Vol.
1560
44%
1580
24%
1600
14%
Results from the "Score" column under the "Text Arena | Coding" Leaderboard tab at https://arena.ai/leaderboard/text/coding-no-style-control with style control off will be used to resolve this market.
The resolution source for this market is the Chatbot Arena LLM Leaderboard found at arena.ai/leaderboard/text. If this resolution source is unavailable at check time, this market will remain open until the leaderboard comes back online and will resolve based on the first check after it becomes available. If permanently unavailable, this market will resolve to "No".
Marché ouvert : Apr 2, 2026, 6:09 PM ET
Resolver
0x65070BE91...Results from the "Score" column under the "Text Arena | Coding" Leaderboard tab at https://arena.ai/leaderboard/text/coding-no-style-control with style control off will be used to resolve this market.
The resolution source for this market is the Chatbot Arena LLM Leaderboard found at arena.ai/leaderboard/text. If this resolution source is unavailable at check time, this market will remain open until the leaderboard comes back online and will resolve based on the first check after it becomes available. If permanently unavailable, this market will resolve to "No".
Resolver
0x65070BE91...Rapid releases from OpenAI and Anthropic, including GPT-5.6 Sol's July 2026 launch and Claude Fable 5's strong agentic coding results, have driven frontier large language models to 95%+ on SWE-Bench Verified and leading Elo ratings in live Coding Arenas. Chinese labs have accelerated competition, with Kimi K3 topping frontend code arenas and DeepSeek V4, GLM-5.2 matching or exceeding proprietary models on agentic benchmarks like Terminal-Bench. This momentum, fueled by iterative model scaling, tool-use improvements, and open-weight gains, supports trader consensus that current leaders or near-term successors will clear elevated arena thresholds by year-end, though benchmark saturation and potential timeline slips remain key uncertainties for resolution.
Résumé expérimental généré par IA à partir des données Polymarket. Ceci n'est pas un conseil de trading et ne joue aucun rôle dans la résolution de ce marché. · Mis à jour



Méfiez-vous des liens externes.
Méfiez-vous des liens externes.
Questions fréquentes