G
GPT-6 Astra
OpenAI's flagship: second on science reasoning at GPQA 95.8, and in the same top price tier as Fable 5.1
Intelligence v5
56.9
Leaderboard rank
#3
Input / output
$10.00/$50.00
Blended 3:1
$20.00
Output speed
81 tok/s
Value
3 pts/$
Capability profile Normalised to the field maximum
Specifications
Vendor
OpenAI
Released
2026-09
Weights
Closed, API only
Context
1,050,000 tokens
Max output
128,000 tokens
Parameters
Undisclosed
Input price
$10.00 / 1M
Output price
$50.00 / 1M
Cache hit
$1.00 / 1M
Output speed
81 tok/s
Intelligence
56.9（v5）
Benchmark scores
How it is measured →Ranked 3 · bars are scaled to the field maximum for that benchmark.
| Benchmarks | Score | Vs. best in field | Gap |
|---|---|---|---|
| AA Intelligence Index v5 Intelligence index | 56.9 | 95% | −3.1 pts |
| AA Intelligence Index v4.3.2 | — | — | not indexed |
| GPQA Diamond Graduate-level science reasoning | 95.8 | 99% | −0.6 pts |
| Terminal-Bench 4.0 Terminal-agent tasks | 59.1 | 88% | −8.1 pts |
| LiveBench Design Document generation (DOCX/PPT/PDF) | 82.0 | 94% | −4.9 pts |
| Vals Index | — | — | not indexed |
Assessment
GPQA Diamond 95.8 is second only to Gemini 4 Argon, and AA Intelligence Index v5 56.9 puts it third. Its $1.00 cache-hit price is the highest of any model here, so long system prompts get expensive.
At the same price
The four models closest on blended price, side by side.
| Model | v5 | Blended | tok/s | Score gap |
|---|---|---|---|---|
| Claude Fable 5.1Anthropic | 49.0 | $20.00 | 57 | -7.9 |
| Claude Opus 5.5Anthropic | 59.1 | $8.00 | 94 | +2.2 |
| Gemini 4 CarbonGoogle | 60.0 | $8.00 | 88 | +3.1 |
| Muse WonderMeta | 44.0 | $8.00 | 143 | -12.9 |
Compare side by side
Line it up against other models on price, speed and every benchmark.
The comparison tool lines up to 4 models side by side with switchable normalisation — built for purchasing decisions.
Open the comparison tool → Back to leaderboard