cd /news/large-language-models/openais-gpt-6-2-astra-takes-3-and-gp… · home › topics › large-language-models › article
[ARTICLE · art-172000] src=cryptobriefing.com ↗ pub= topic=large-language-models verified=true sentiment=↑ positive

OpenAI’s GPT-6.2 Astra takes #3 and GPT-6.2 Sol lands #7 on Code Arena WebDev leaderboard

OpenAI's GPT-6.2 Astra (Max) debuted at #3 with 1781 points and GPT-6.2 Sol (Max) at #7 with 1757 points on Code Arena's WebDev leaderboard as of November 18, 2026, one day after the November 17, 2026 launch. Astra's median cost per task fell 42% to $5.45 from $9.46 while gaining 18 points over GPT-6 Astra (Max), and Sol's cost dropped 21% to $1.45 from $1.84 with a 13-point gain over GPT-6.1 Sol (Max); OpenAI also cut Astra's API price 20% to $8 per million tokens, with Sol unchanged at $2 per million tokens. Both models sit on the performance-cost Pareto frontier, though Astra's agent scores remain pending despite OpenAI's agentic positioning, and Astra ranks #2 on the Artificial Analysis Intelligence Index.

recommendation

Route cheaper tasks to GPT-6.2 Sol at $1.45 median per task and reserve Astra's 24-point-higher score for work that justifies its roughly 3.8x cost.

more in playbook: Controlling LLM cost
by read2 min views8 publishedNov 18, 2026
OpenAI’s GPT-6.2 Astra takes #3 and GPT-6.2 Sol lands #7 on Code Arena WebDev leaderboard
Image: Cryptobriefing (auto-discovered)

OpenAI / Wikimedia Commons (Public domain)

OpenAI's new model pair climbed the crowdsourced rankings while cutting median costs per task, though agent scores are still pending

OpenAI has two new models out, and both landed in the top ten of Code Arena’s WebDev leaderboard. GPT-6.2 Astra (Max) took the #3 spot with 1781 points, while GPT-6.2 Sol (Max) came in at #7 with 1757 points.

OpenAI unveiled GPT-6.2 Astra and GPT-6.2 Sol on November 17, 2026. The company pitched the pair as built for stronger real-world performance and agentic work, meaning tasks where the AI takes multi-step actions instead of answering a single prompt.

The leaderboard placements come from Code Arena, a platform run by Arena.ai that builds its rankings from crowdsourced, community-driven evaluations. The WebDev standings cited here reflect the board as of November 18, 2026, one day after launch.

Astra (Max) posted 1781 points, an improvement of 18 points over GPT-6 Astra (Max). The cost side is not modest. Astra’s median cost per task fell from $9.46 to $5.45, a 42% reduction.

Sol (Max) followed a similar pattern on a smaller scale. It scored 13 points higher than GPT-6.1 Sol (Max), while its median cost per task dropped from $1.84 to $1.45, a 21% decrease.

Both models landed on what’s known as the performance-cost Pareto frontier. A model sits on that frontier when no other model on the board delivers a higher score without also costing more.

AI, tech, and the markets they move—in one daily briefing.

Daily. Free. Join 34,000+ readers across crypto, finance, and policy.

Astra beats Sol by 24 points on the WebDev board. It also costs roughly 3.8 times as much per task, based on the median figures.

Astra did pull off one notable category win. In the Consumer Product category, it climbed from #4 under prior versions to #1.

API pricing tells a similar story. OpenAI cut Astra’s price by 20% to $8 per million tokens, while Sol stays at its previous rate of $2 per million tokens.

The broader benchmark picture is more mixed than the Code Arena headline numbers suggest. Astra ranks #2 on the Artificial Analysis Intelligence Index, a strong showing on a general measure of model capability. However, it ran into difficulties on certain coding indices.

There’s also an asterisk on the agentic pitch. OpenAI built these models with agentic tasks in mind, but agent scores are still pending.

Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

── more in #large-language-models 4 stories · sorted by recency
── more on @openai 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
→ Live at https://your-agent.zahid.host ✓
Get free account → Pricing
from €0/mo · no card required
LIVE [news/openais-gpt-6-2-astr…] indexed:0 read:2min 2026-11-18 · —