ACB
64.5 score
LLM Leaderboard · Proprietary
GPT-6 Astra benchmarks — leaderboard rank, pricing, and comparisons.
GPT-6 Astra ranks #8 on the AskClash LLM leaderboard with an overall score of 70.8. It scores 67.0 on the Coding Agent Index and 59.1 on Terminal-Bench. Compare GPT-6 Astra vs GPT, Claude, Gemini, DeepSeek, open-weight, and frontier AI models using public benchmark scores, token pricing, context window, and access details.
Rank #8AskClash overall score: 70.8
$10.0 / $50.0Input and output token price, when published. Context: 1.05M.
Visit websiteVisit the model provider's website.
GPT-6 Astra benchmark snapshot
AskClash combines public LLM benchmark cells into a weighted percentile score and penalizes missing coverage so narrow rows do not dominate better-measured models.
Overall70.8
Benchmark cells13
Context1.05M
CreatorOpenAI
GPT-6 Astra public benchmark scores
Cached benchmark values can include HLE, GPQA, SWE-bench, SWE-Pro, SWE-Atlas, Terminal-Bench, MCP Atlas, MMMU-Pro, ARC-AGI-2, Tau2, and model-specific coding or agent scores.
RWT
9.5 score
Coding Agent Index
67.0 score
HLE
57.2 score
GPQA
96.0 score
Terminal-Bench
59.1 score
DeepSWE
73.2 score
GDPval-AA
1629.3 score
Finance Agent
53.5 score
MMMU-Pro
87.0 score
ARC-AGI 2
95.0 score
MRCR
100.0 score
GPT-6 Astra vs other AI models
Use these comparison links to evaluate GPT-6 Astra against nearby LLMs by benchmark score, price, context window, and provider.
Related AI and tech coverage
Cached AskClash article matches that can provide release, provider, benchmark, pricing, or market context around this model.
GPT-6-Astra solves 49 of 50 physics research problems at $2.40 apiece
On the TheoPhysEval benchmark, GPT-6-Astra solves 49 of 50 unpublished theoretical physics calculations as accurately as leading specialists. In some cases where the agent seemed to fail, the experts' own reference solutions turned out to contain errors. But t
Meta claims new Muse Wonder AI models rival OpenAI, Anthropic
Meta’s new models are its most expensive to date, with the standard Wonder model charging $4 per million input tokens, matching Claude Opus 5.5 but undercutting GPT-6 Astra and Claude Fable 5.1, which charge $10 and $50 respectively. Bindu Reddy, CEO of Abacus
web-edit 0.1a1
Release: web-edit 0.1a1 Simon Willison’s Weblog Newsletter Sponsored by: Pomerium — Static OAuth scopes won't stop AI agent breakouts. Upgrade to Pomerium's agentic access management trusted by NVIDIA and OpenAI 13th October 2026 Release web-edit 0.1a1 — Web a
web-edit 0.1a0
Release: web-edit 0.1a0 Simon Willison’s Weblog Newsletter Sponsored by: Pomerium — Static OAuth scopes won't stop AI agent breakouts. Upgrade to Pomerium's agentic access management trusted by NVIDIA and OpenAI 12th October 2026 Release web-edit 0.1a0 — Web a
Last cached leaderboard date: . This model page is generated from the AskClash LLM Leaderboard cache and linked from the live leaderboard.