Skip to content
Tier A — Frontier
Runs in:USMade in:United States
OpenAI

gpt-6-astra

Tier A — Frontier · 1.05M tokens

Tokonomix Editorial Team·Reviewed by Mes Kalkan··
Section 01

Speed analysis

Latency measured across all benchmark runs. P50 (median) and P95 (95th percentile) give a realistic picture of response speed under normal and peak load.

P50 latency (median)P95 latency107 runs
882274446066467832910-2611-22ms
Section 02

Quality scores

How this model compares to the rest of the field on each prompt category, from a pairwise fit over the same prompts. The raw judge score sits underneath each number.

78%
Coding
judge mean 91
80%
Creative
judge mean 81
75%
Factual
judge mean 75
52%
Reasoning
judge mean 95

Win rate per category: how often this model beats a field-average model on a prompt from that category. 50% is average, not a failing grade. It is not a percentage of correct answers.

Section 03

Pricing

What you pay per million tokens when you use this model on Tokonomix, plus an estimate for a typical conversation.

💰
API rates — gpt-6-astra
$13.00 per 1M input tokens
$65.00 per 1M output tokens
≈ $0.0208 per typical conversation (800 tokens)
Input vs output price (per 1M tokens)
per 1M input tokens$13.00
per 1M output tokens$65.00
Section 04

Tokens per second

Throughput in tokens per second, derived from measured P50 latency. Higher is better; fluctuations track provider-side load.

Throughput (tokens / s)204 / avg 135
22522

Estimated from P50 latency × 200 output tokens — the absolute number depends on this assumption; the trend is what matters.

Section 05

Capabilities

toolssource: seed-newest-generation-2026-09-30visionjson modepdf inputreasoningjson schemaparallel toolsprompt cachingmax output tokens: 128000
Section 06

Availability

Availability

How often this model answers when we call it — measured across real API requests and live tests over the last 30 days. This is separate from quality: these numbers only tell you whether the model responds, not how good the answer is.

Last 7 days

100.0%

n=1

Last 30 days

100.0%

n=1

Median response time

2,195ms

n=1

Based on 375 measurements over the last 30 days.

Technical details

Only live API calls and live-test requests count — internal probes and benchmark runs are excluded.

Calls with a custom API key (BYOK) are excluded: those failures are key-specific, not a sign of model downtime.

Failed calls are NOT included in quality scores — quality is measured on successful responses only. Availability and quality are independent signals.

Median response time (p50) across successful calls with a recorded duration. Outliers (very slow or very fast calls) pull the median less than the average.

Total calls (30d)

1

OK responses (30d)

1

Total calls (7d)

1

OK responses (7d)

1

Section 07

Tokonomix benchmark verdicts

⚖️
Endorsed by 1 judge
Independent LLM judges evaluated this model on our weekly intelligence tests
claude-sonnet-4-586/100 · 28 runs
22 correct3 partial3 wrong79% accuracy
● 2026-11-22

GPT-6-Astra shows major quality leap with 32-point improvement

GPT-6-Astra has demonstrated substantial improvement in its second benchmark window, with overall quality jumping from 59 to 91 points. The model achieved a perfect 100 score in coding tasks, up from 74, indicating significant refinement in code generation and problem-solving capabilities. Creative performance reached 81 points in the current window, though direct comparison is limited by the absence of this metric in previous testing. Reasoning remained exceptionally stable at 92 points across both windows, suggesting this was already a strength. Notably, the previous factual category score of 11 is not present in current results, making it unclear whether this weakness was addressed or simply not retested. Latency improvements are equally impressive, with response times dropping 41 percent from 6389ms to 3764ms at the median. However, the current window includes only 3 test runs compared to 4 previously, which may affect statistical confidence. Users can expect substantially faster responses and notably stronger coding assistance, while reasoning capabilities continue to perform at their established high level. The model appears to have undergone significant optimization between benchmark periods.

Quality

91.0

Latency p50

3,764 ms

Test runs

3

✓ Quality improved 32 points✓ Perfect coding score achieved✓ Latency reduced 41 percent✗ Fewer test runs this window
Last automated test
Nov 22, 2026 · 08:04 UTC · Speed benchmark
P50 latency
980 ms
P95 latency
1207 ms
Errors
0 / 6 runs
Last reviewed by Tokonomix Team·November 22, 2026