OpenAI

GPT-6 Sol (low)

GPT-6 Sol running at low reasoning effort.

#168released Aug 20, 2026·Proprietary·Not disclosed
textimage400K ctxreasoningeffort: low
API providersGet API Key →Leaderboard
Intelligence
70.0 #168
Top 22% tier
Output Speed
250 t/s
TTFT 280 ms · ITL 4.0 ms/tok
Blended price (3:1)
$1.667 /1M
In $1/M · Out $4/M
Context
400K
≈ 1,000 A4 pages · max out 32,768

Quality Domain Radar

6-Axis Frontier Evaluation vs Category Average & Flagship

GPT-6 Sol (low)
Claude Mythos Preview (#1)
Average (68)
Coding (50)Mathematics (75)Hard Science (GPQA) (70)Agentic & Tool Use (68)Knowledge (MMLU) (45)Vision / Multimodal (30)

Radar axes are approximated from category scores where direct benchmark measurements are missing — see Methodology.

Strengths & Trade-offs

Where it wins

  • Extremely fast — top-3 output speed on the leaderboard
  • Very low price for an OpenAI frontier-tier model
  • Great value tier for agentic loops

Trade-offs to watch

  • Shallower reasoning than Astra/Luna
  • 400K context is small vs. 1M+ peers

Killer Use Cases

Voice & realtime agents
Sub-second token loops for conversational pipelines.
Agent workers
Cheap fast sub-agents under a frontier orchestrator.
Inline completions
IDE-style latency-sensitive calls.

Developer Quickstart & API

Set OPENAI_API_KEY first — get an API key
curl https://api.openai.com/v1/responses \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "content-type: application/json" \
  -d '{
    "model": "gpt-6-sol-low",
    "input": "Refactor this function for O(n) time and explain the change."
  }'

Official Resources & Links

API deployment

Compare hosts & latency

Throughput, TTFT and cache pricing across Groq, Cerebras, Together, DeepInfra, Fireworks, Bedrock and Azure.

GroqCerebrasTogether AIDeepInfraFireworks AIAWS BedrockAzure AI
View providers matrix →

Scores by Source

Live Specs & Architecture

Under the hood
GPT-6 Sol — speed-first tier
  • Throughput-optimized GPT-6 tier at $1/$4
  • 200+ tok/s output for interactive agents
  • 400K context window
Release date
Aug 20, 2026
Architecture
GPT-6 Sol — speed-first tier
Parameters
Not disclosed
License
Proprietary
Context Window
400,000 tokens
Max Output
32,768 tokens
Modalities
text + image
Cache Pricing
read $0.1 · write $1.25 /1M
Reasoning
Effort levels
Input Price
$1/M
Output Price
$4/M
Output Speed
250 t/s
Time to First Token
280 ms
Data Updated
Sep 26, 2026

Estimated monthly bill

Estimated monthly spend
$0.00
Calculating savings vs GPT-4o...

Cite & embed

GitHub Markdown badge
Loading badge markdown...
BibTeX
@misc{llmpodium2026gpt6sollow,
  title = {Benchmark Analysis and Intelligence Evaluation of GPT-6 Sol (low)},
  author = {LLMPodium Research},
  year = {2026},
  url = {https://llmpodium.com/models/gpt-6-sol-low}
}

Embed Score Badge

Add a live Podium Score badge to your GitHub or Hugging Face README — it links back to the live ranking.

Preview
GPT-6 Sol (low) LLMPodium Score
Markdown
Loading badge snippet...
HTML
Loading badge snippet...

More from OpenAI

Alternative Models

Closest alternatives to GPT-6 Sol (low) based on capabilities, pricing, and benchmark scores.

0 / 4