OpenAI · launched 17 Nov 2026
GPT-6.2 Sol
OpenAI's everyday model for work from coding to powering agents at scale.
Read the official announcement- Vendor
- OpenAI
- Launch date
- 17 Nov 2026
- Input price
- $2.00 / 1M tokens
- Output price
- $10 / 1M tokens
- Cached input
- $0.10 / 1M tokens
- Batch discount
- 50%
- Context window
- 1.05M tokens
- Max output
- 128K tokens
- Input types
- Text, Image
- API model ID
- gpt-6.2-sol
Listed rates apply up to 272K input tokens. Above 272K, input costs $4, cached input $0.20 and output $15 per million tokens for the full request.
Official pricing source · checked 18 Nov 2026
Input price
$2.00
was $2.00 · Same price
Output price
$10
was $10 · Same price
Blended price
$4.00
was $4.00 · Same price
Listed rates apply up to 272K input tokens. Above 272K, input costs $4, cached input $0.20 and output $15 per million tokens for the full request.
- Agents' Last Exam v1.1max effort · $1.32 per task · reported by OpenAI73.04%
- Terminal-Bench Science 0.1max effort · $2.21 per task · reported by OpenAI68%
- Vals Science Index v1xhigh effort · $0.71 per task · reported by OpenAI58.5%
- FrontierCode 2.0max effort · $1.29 per task · reported by OpenAI43.8%
- LifeValmax effort · $1.27 per task · reported by OpenAI83.8%
- Harvey-LAB-AA v1.1max effort · $1.92 per task · reported by OpenAI11.1%
- BenchCAD 2.1 Coremax effort · $1.60 per task · reported by OpenAI73.1%
- AutomationBench 1.0.6high effort · $0.21 per task · reported by OpenAI41.2%
- SecureSWEmax effort · $0.99 per task · reported by OpenAI55.6%
- PatchBenchmax effort · $0.85 per task · reported by OpenAI75.6%
Price compared with other models
| Model | Launched | Input / 1M | Output / 1M | Cached / 1M | Blended / 1M | vs GPT-6.2 Sol |
|---|---|---|---|---|---|---|
| Claude Haiku 5.5AnthropicListed prices and comparisons apply to prompts up to 100K tokens. Above 100K, input costs $0.50, output $2.50 and cache reads $0.05 per million tokens. | 7 Oct 2026 | $0.10 | $0.50 | $0.01 | $0.20 | 95% cheaper |
| GPT-6 LunaOpenAI | 22 Sept 2026 | $0.10 | $0.50 | $0.01 | $0.20 | 95% cheaper |
| GPT-5.6 LunaOpenAI | 9 Jul 2026 | $0.20 | $1.20 | $0.02 | $0.45 | 89% cheaper |
| Gemini 3.5 Flash-LiteGoogle DeepMindText, image, video and audio inputs share the listed rate. Cache storage and grounding are billed separately. | 21 Jul 2026 | $0.30 | $2.50 | $0.03 | $0.85 | 79% cheaper |
| Gemini 3.8 FlashGoogle DeepMindIntroductory rates through 31 December 2026. From 1 January 2027, input is $1.50, output $7.50 and cached input $0.15 per million tokens. Cache storage and grounding are billed separately. | 2 Sept 2026 | $0.75 | $3.75 | $0.075 | $1.50 | 63% cheaper |
| Claude Haiku 4.5Anthropic | 15 Oct 2025 | $1.00 | $5.00 | $0.10 | $2.00 | 50% cheaper |
| GPT-6.1 TerraOpenAIListed rates apply up to 272K input tokens. Above 272K, input costs $2, cached input $0.10 and output $7.50 per million tokens for the full request. | 26 Oct 2026 | $1.00 | $5.00 | $0.05 | $2.00 | 50% cheaper |
| Grok 4.7SpaceXAIBase rates apply up to 200K context tokens; higher-context requests use different rates. Batch API is not supported. The separate fast variant costs twice as much. | 21 Sept 2026 | $2.00 | $6.00 | $0.50 | $3.00 | 25% cheaper |
| Claude Sonnet 5Anthropic | 30 Jun 2026 | $2.00 | $10 | $0.20 | $4.00 | Same price |
| Claude Sonnet 5.5AnthropicCache reads were reduced from $0.20 to $0.10 per million tokens on 7 October 2026. | 28 Sept 2026 | $2.00 | $10 | $0.10 | $4.00 | Same price |
| Gemini 4 ArgonGoogle DeepMindIntroductory rates through 30 November 2026. From 1 December 2026, input is $4, output $20 and cached input $0.20 per million tokens. Caching is implicit only. Long decode raises the output limit to 1,048,576 tokens. | 30 Sept 2026 | $2.00 | $10 | $0.10 | $4.00 | Same price |
| GPT-6 SolOpenAIListed rates apply up to 272K input tokens. Above 272K, input costs $4, cached input $0.40 and output $15 per million tokens for the full request. | 22 Sept 2026 | $2.00 | $10 | $0.20 | $4.00 | Same price |
| GPT-6.1 SolOpenAI · previous generationListed rates apply up to 272K input tokens. Above 272K, input costs $4, cached input $0.20 and output $15 per million tokens for the full request. | 29 Sept 2026 | $2.00 | $10 | $0.10 | $4.00 | Same price |
| GPT-6.2 SolOpenAIListed rates apply up to 272K input tokens. Above 272K, input costs $4, cached input $0.20 and output $15 per million tokens for the full request. | 17 Nov 2026 | $2.00 | $10 | $0.10 | $4.00 | Baseline |
| Gemini 3.1 ProGoogle DeepMind · PreviewRates apply to prompts up to 200K tokens. Above 200K, input is $4, output $18 and cached input $0.40 per million tokens. Cache storage and grounding are billed separately. | 19 Feb 2026 | $2.00 | $12 | $0.20 | $4.50 | 13% more |
| GPT-5.6 TerraOpenAIOpenAI's model documentation supports the Batch API but does not list a batch price. | 9 Jul 2026 | $2.00 | $12 | $0.20 | $4.50 | 13% more |
| Claude Opus 5.5Anthropic | 22 Sept 2026 | $4.00 | $20 | $0.20 | $8.00 | 2× the price |
| GPT-5.6 SolOpenAIPrice before the GPT-6 launch, from OpenAI's announcement. OpenRouter now lists $2 / $10. | 9 Jul 2026 | $4.00 | $20 | – | $8.00 | 2× the price |
| Claude Opus 5Anthropic | 24 Jul 2026 | $5.00 | $25 | $0.50 | $10 | 2.5× the price |
| Claude Fable 5.2Anthropic | 9 Nov 2026 | $8.00 | $40 | $0.20 | $16 | 4× the price |
| GPT-6.2 AstraOpenAIListed rates apply up to 272K input tokens. Above 272K, input costs $16, cached input $0.80 and output $60 per million tokens for the full request. | 17 Nov 2026 | $8.00 | $40 | $0.40 | $16 | 4× the price |
| Claude Fable 5Anthropic | 9 Jun 2026 | $10 | $50 | $1.00 | $20 | 5× the price |
| Claude Fable 5.1Anthropic | 1 Sept 2026 | $10 | $50 | $0.25 | $20 | 5× the price |
| GPT-6 AstraOpenAI | 4 Sept 2026 | $10 | $50 | $1.00 | $20 | 5× the price |
GPT-6.1 Sol and GPT-6 Sol rates were checked on 8 October 2026 against OpenAI model documentation. USD per million tokens, checked 2 Oct 2026 via OpenRouter. Comparison uses the blended price (3 input tokens for every output token). Haiku 5.5 and Sonnet 5.5 prices were updated from Anthropic's 7 October 2026 announcement. Gemini and Grok prices were rechecked against official Google and xAI documentation on 27 October 2026. GPT-6.1 Terra prices were checked against OpenAI documentation on 27 October 2026. Claude Fable 5.2 prices are from Anthropic's 9 November 2026 announcement.
Official benchmark results
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Long-horizon professional workflows completed by agents.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 65.77% | $0.47 |
| GPT-6.2 Astra | medium | 72.06% | $1.06 |
| GPT-6.2 Astra | high | 75.96% | $2.05 |
| GPT-6.2 Astra | xhigh | 78.33% | $2.89 |
| GPT-6.2 Astra | max | 77.61% | $4.47 |
| GPT-6.2 Sol | low | 66.65% | $0.27 |
| GPT-6.2 Sol | medium | 70.26% | $0.57 |
| GPT-6.2 Sol | high | 72.04% | $0.69 |
| GPT-6.2 Sol | xhigh | 72.48% | $0.86 |
| GPT-6.2 Sol | max | 73.04% | $1.32 |
| GPT-6 Astra | low | 67.87% | $5.40 |
| GPT-6 Astra | medium | 71.16% | $6.21 |
| GPT-6 Astra | high | 71.04% | $6.87 |
| GPT-6 Astra | xhigh | 71.68% | $7.56 |
| GPT-6 Astra | max | 73.14% | $8.99 |
| GPT-6.1 Sol | low | 60.98% | $0.21 |
| GPT-6.1 Sol | medium | 68.77% | $0.46 |
| GPT-6.1 Sol | high | 72.34% | $0.98 |
| GPT-6.1 Sol | xhigh | 71.92% | $0.98 |
| GPT-6.1 Sol | max | 72.45% | $2.02 |
| Claude Fable 5.2 | max | 72.79% | $60.1 |
| Claude Opus 5.5 | max | 71.6% | $17.6 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
- Claude Fable 5.2 leaderboard results exclude nine tasks that used a fallback model, as reported by OpenAI.
- OpenAI's chart also shows Gemini 4 Carbon, which is not in the tracker, so its results are left out.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Scientific research workflows carried out in a terminal.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2 (mini-swe-agent harness)
- Claude Opus 5.5
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 56.3% | $2.50 |
| GPT-6.2 Astra | medium | 72.3% | $4.20 |
| GPT-6.2 Astra | high | 74.9% | $5.20 |
| GPT-6.2 Astra | xhigh | 74% | $5.37 |
| GPT-6.2 Astra | max | 77.1% | $7.15 |
| GPT-6.2 Sol | low | 43.4% | $0.86 |
| GPT-6.2 Sol | medium | 63.1% | $1.34 |
| GPT-6.2 Sol | high | 67.7% | $1.51 |
| GPT-6.2 Sol | xhigh | 64.3% | $1.61 |
| GPT-6.2 Sol | max | 68% | $2.21 |
| GPT-6 Astra | low | 55.4% | $11.4 |
| GPT-6 Astra | medium | 57.4% | $12.3 |
| GPT-6 Astra | high | 62% | $14.9 |
| GPT-6 Astra | xhigh | 60.9% | $15.8 |
| GPT-6 Astra | max | 68.1% | $23.8 |
| GPT-6.1 Sol | low | 43.7% | $1.79 |
| GPT-6.1 Sol | medium | 47.6% | $2.34 |
| GPT-6.1 Sol | high | 51.1% | $2.76 |
| GPT-6.1 Sol | xhigh | 53.7% | $2.89 |
| GPT-6.1 Sol | max | 57% | $5.47 |
| Claude Fable 5.2 (mini-swe-agent harness) | low | 53.8% | $8.78 |
| Claude Fable 5.2 (mini-swe-agent harness) | medium | 60.5% | $11 |
| Claude Fable 5.2 (mini-swe-agent harness) | high | 70% | $24.3 |
| Claude Fable 5.2 (mini-swe-agent harness) | xhigh | 67.6% | $31.4 |
| Claude Fable 5.2 (mini-swe-agent harness) | max | 73.3% | $45.2 |
| Claude Opus 5.5 | max | 63.3% | $23.2 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
- Claude Fable 5.2 was run in the mini-swe-agent harness, as reported by OpenAI.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Interpreting real biological data and running experiments to infer how a simulated system works.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Accuracy (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 59.5% | $1.05 |
| GPT-6.2 Astra | medium | 67.6% | $1.88 |
| GPT-6.2 Astra | high | 68.1% | $2.25 |
| GPT-6.2 Astra | xhigh | 64.8% | $2.47 |
| GPT-6.2 Astra | max | 68.4% | $4.17 |
| GPT-6.2 Sol | low | 46.3% | $0.19 |
| GPT-6.2 Sol | medium | 56.5% | $0.44 |
| GPT-6.2 Sol | high | 57.6% | $0.63 |
| GPT-6.2 Sol | xhigh | 58.5% | $0.71 |
| GPT-6.2 Sol | max | 57.8% | $1.03 |
| GPT-6 Astra | max | 65.1% | $8.03 |
| GPT-6.1 Sol | max | 59.6% | $1.36 |
| Claude Fable 5.2 | high | 63.74% | $8.79 |
| Claude Opus 5.5 | max | 58.7% | $8.71 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Cognition's benchmark of whether changes to real codebases meet maintainers' standards for merging.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 38.1% | $1.41 |
| GPT-6.2 Astra | medium | 42.2% | $2.47 |
| GPT-6.2 Astra | high | 43.6% | $3.95 |
| GPT-6.2 Astra | xhigh | 44.1% | $4.29 |
| GPT-6.2 Astra | max | 44.2% | $5.97 |
| GPT-6.2 Sol | low | 33.1% | $0.40 |
| GPT-6.2 Sol | medium | 38.3% | $0.68 |
| GPT-6.2 Sol | high | 39.7% | $0.88 |
| GPT-6.2 Sol | xhigh | 39.8% | $0.98 |
| GPT-6.2 Sol | max | 43.8% | $1.29 |
| GPT-6 Astra | low | 31.1% | $2.44 |
| GPT-6 Astra | medium | 36.8% | $3.49 |
| GPT-6 Astra | high | 38.3% | $4.28 |
| GPT-6 Astra | xhigh | 41.4% | $4.54 |
| GPT-6 Astra | max | 42.8% | $6.68 |
| GPT-6.1 Sol | low | 25.6% | $0.34 |
| GPT-6.1 Sol | medium | 29.6% | $0.49 |
| GPT-6.1 Sol | high | 34.8% | $0.65 |
| GPT-6.1 Sol | xhigh | 32.3% | $0.70 |
| GPT-6.1 Sol | max | 39% | $1.03 |
| Claude Fable 5.2 | low | 34.2% | $2.10 |
| Claude Fable 5.2 | medium | 37.59% | $2.48 |
| Claude Fable 5.2 | high | 45.29% | $6.72 |
| Claude Fable 5.2 | xhigh | 49.11% | $14.2 |
| Claude Fable 5.2 | max | 45.65% | $44.7 |
| Claude Opus 5.5 | low | 14.7% | $0.75 |
| Claude Opus 5.5 | medium | 30.1% | $1.70 |
| Claude Opus 5.5 | high | 34.5% | $2.42 |
| Claude Opus 5.5 | xhigh | 43.4% | $5.18 |
| Claude Opus 5.5 | max | 45.5% | $12.8 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
- OpenAI's chart also shows Gemini 4 Carbon, which is not in the tracker, so its results are left out.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Everyday tasks tailored to someone's needs, such as finding a home that fits a family's budget and commute.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 46.3% | $0.59 |
| GPT-6.2 Astra | medium | 78.9% | $1.56 |
| GPT-6.2 Astra | high | 83.1% | $2.41 |
| GPT-6.2 Astra | xhigh | 83.7% | $2.94 |
| GPT-6.2 Astra | max | 86.1% | $4.23 |
| GPT-6.2 Sol | low | 25.3% | $0.089 |
| GPT-6.2 Sol | medium | 70.3% | $0.38 |
| GPT-6.2 Sol | high | 80.8% | $0.65 |
| GPT-6.2 Sol | xhigh | 82.6% | $0.88 |
| GPT-6.2 Sol | max | 83.8% | $1.27 |
| GPT-6 Astra | low | 67.5% | $3.51 |
| GPT-6 Astra | medium | 74.3% | $5.30 |
| GPT-6 Astra | high | 76% | $6.52 |
| GPT-6 Astra | xhigh | 77.5% | $6.94 |
| GPT-6 Astra | max | 79.8% | $9.04 |
| GPT-6.1 Sol | low | 66.8% | $0.52 |
| GPT-6.1 Sol | medium | 75.5% | $0.73 |
| GPT-6.1 Sol | high | 77.1% | $0.88 |
| GPT-6.1 Sol | xhigh | 77.8% | $0.96 |
| GPT-6.1 Sol | max | 79.5% | $1.27 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Legal work tasks, scored by the all-pass rate gated on hallucinations.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Hallucination-gated all-pass rate (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 2.2% | $0.86 |
| GPT-6.2 Astra | medium | 8.1% | $2.49 |
| GPT-6.2 Astra | high | 13.3% | $3.86 |
| GPT-6.2 Astra | xhigh | 12.8% | $4.31 |
| GPT-6.2 Astra | max | 15% | $6.35 |
| GPT-6.2 Sol | low | 3.1% | $0.25 |
| GPT-6.2 Sol | medium | 5% | $0.71 |
| GPT-6.2 Sol | high | 5.8% | $1.03 |
| GPT-6.2 Sol | xhigh | 6.1% | $1.23 |
| GPT-6.2 Sol | max | 11.1% | $1.92 |
| GPT-6 Astra | max | 8.6% | $13.7 |
| GPT-6.1 Sol | max | 6.9% | $2.48 |
| Claude Fable 5.2 | low | 4.2% | $9.16 |
| Claude Fable 5.2 | medium | 6.4% | $9.67 |
| Claude Fable 5.2 | high | 4.7% | $15.9 |
| Claude Fable 5.2 | xhigh | 3.9% | $24.3 |
| Claude Fable 5.2 | max | 7.2% | $39.8 |
| Claude Opus 5.5 | max | 4.2% | $17.9 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Computer-aided design tasks.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 42.6% | $0.98 |
| GPT-6.2 Astra | medium | 68.5% | $1.94 |
| GPT-6.2 Astra | high | 74.6% | $2.89 |
| GPT-6.2 Astra | xhigh | 77.6% | $3.36 |
| GPT-6.2 Astra | max | 82.5% | $5.43 |
| GPT-6.2 Sol | low | 23.9% | $0.15 |
| GPT-6.2 Sol | medium | 41.2% | $0.29 |
| GPT-6.2 Sol | high | 64% | $0.75 |
| GPT-6.2 Sol | xhigh | 70.4% | $1.00 |
| GPT-6.2 Sol | max | 73.1% | $1.60 |
| GPT-6 Astra | low | 20.3% | $1.14 |
| GPT-6 Astra | medium | 38.7% | $3.73 |
| GPT-6 Astra | high | 51.1% | $5.38 |
| GPT-6 Astra | xhigh | 53.9% | $6.32 |
| GPT-6 Astra | max | 62.6% | $9.71 |
| GPT-6.1 Sol | low | 13.5% | $0.19 |
| GPT-6.1 Sol | medium | 31.2% | $0.55 |
| GPT-6.1 Sol | high | 42.1% | $0.87 |
| GPT-6.1 Sol | xhigh | 45.6% | $1.05 |
| GPT-6.1 Sol | max | 52.6% | $1.61 |
| Claude Fable 5.2 | low | 26.9% | $1.89 |
| Claude Fable 5.2 | medium | 32.4% | $2.66 |
| Claude Fable 5.2 | high | 61% | $9.99 |
| Claude Fable 5.2 | xhigh | 74.1% | $23.2 |
| Claude Fable 5.2 | max | 83.5% | $50.8 |
| Claude Opus 5.5 | low | 6.3% | $0.17 |
| Claude Opus 5.5 | medium | 19.1% | $1.20 |
| Claude Opus 5.5 | high | 29.7% | $2.70 |
| Claude Opus 5.5 | xhigh | 54.5% | $9.00 |
| Claude Opus 5.5 | max | 65.3% | $20 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Workflow automation tasks.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
- Claude Fable 5.2
- Claude Opus 5.5
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 37.1% | $0.44 |
| GPT-6.2 Astra | medium | 43.8% | $0.55 |
| GPT-6.2 Astra | high | 47.9% | $0.67 |
| GPT-6.2 Astra | xhigh | 47.2% | $0.70 |
| GPT-6.2 Astra | max | 52.4% | $0.83 |
| GPT-6.2 Sol | low | 31.2% | $0.13 |
| GPT-6.2 Sol | medium | 40.2% | $0.18 |
| GPT-6.2 Sol | high | 41.2% | $0.21 |
| GPT-6.2 Sol | xhigh | 40.6% | $0.23 |
| GPT-6.2 Sol | max | 41.1% | $0.27 |
| GPT-6 Astra | low | 30.3% | $1.08 |
| GPT-6 Astra | medium | 34.1% | $1.27 |
| GPT-6 Astra | high | 37.1% | $1.44 |
| GPT-6 Astra | xhigh | 39% | $1.50 |
| GPT-6 Astra | max | 41.4% | $1.73 |
| GPT-6.1 Sol | low | 24.7% | $0.16 |
| GPT-6.1 Sol | medium | 31.7% | $0.19 |
| GPT-6.1 Sol | high | 33.2% | $0.23 |
| GPT-6.1 Sol | xhigh | 35.5% | $0.25 |
| GPT-6.1 Sol | max | 36.1% | $0.30 |
| Claude Fable 5.2 | low | 39.73% | $1.30 |
| Claude Fable 5.2 | medium | 40.18% | $1.35 |
| Claude Fable 5.2 | high | 44.29% | $1.69 |
| Claude Fable 5.2 | xhigh | 47.95% | $2.29 |
| Claude Fable 5.2 | max | 51.45% | $4.64 |
| Claude Opus 5.5 | low | 24.2% | $0.51 |
| Claude Opus 5.5 | medium | 29.5% | $0.65 |
| Claude Opus 5.5 | high | 33% | $0.71 |
| Claude Opus 5.5 | xhigh | 35.8% | $0.89 |
| Claude Opus 5.5 | max | 42.5% | $1.44 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
- OpenAI's chart also shows Gemini 4 Carbon, which is not in the tracker, so its results are left out.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
OpenAI's internal benchmark of writing functional, secure code for security-sensitive components such as serving files or handling database queries.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 54.5% | $0.78 |
| GPT-6.2 Astra | medium | 57.7% | $1.40 |
| GPT-6.2 Astra | high | 61.7% | $2.50 |
| GPT-6.2 Astra | xhigh | 63.8% | $3.12 |
| GPT-6.2 Astra | max | 65.7% | $5.26 |
| GPT-6.2 Sol | low | 43.1% | $0.16 |
| GPT-6.2 Sol | medium | 47.1% | $0.25 |
| GPT-6.2 Sol | high | 51.9% | $0.45 |
| GPT-6.2 Sol | xhigh | 54% | $0.62 |
| GPT-6.2 Sol | max | 55.6% | $0.99 |
| GPT-6 Astra | low | 41.3% | $1.10 |
| GPT-6 Astra | medium | 45.7% | $1.91 |
| GPT-6 Astra | high | 47.9% | $2.88 |
| GPT-6 Astra | xhigh | 48.7% | $3.21 |
| GPT-6 Astra | max | 46.5% | $4.91 |
| GPT-6.1 Sol | low | 36.2% | $0.13 |
| GPT-6.1 Sol | medium | 46.5% | $0.30 |
| GPT-6.1 Sol | high | 44.4% | $0.47 |
| GPT-6.1 Sol | xhigh | 46% | $0.58 |
| GPT-6.1 Sol | max | 49.7% | $0.87 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.
Score against cost · reported by OpenAI on 17 Nov 2026 · exact values
Producing security patches that preserve intended behavior.
- GPT-6.2 Astra
- GPT-6.2 Sol
- GPT-6 Astra
- GPT-6.1 Sol
Show data table
| Model | Effort | Score (%) | Cost per task |
|---|---|---|---|
| GPT-6.2 Astra | low | 71.1% | $0.38 |
| GPT-6.2 Astra | medium | 75.6% | $0.85 |
| GPT-6.2 Astra | high | 78.8% | $1.79 |
| GPT-6.2 Astra | xhigh | 79.7% | $2.45 |
| GPT-6.2 Astra | max | 79.7% | $4.18 |
| GPT-6.2 Sol | low | 65.6% | $0.085 |
| GPT-6.2 Sol | medium | 70.5% | $0.17 |
| GPT-6.2 Sol | high | 74.1% | $0.36 |
| GPT-6.2 Sol | xhigh | 74.2% | $0.51 |
| GPT-6.2 Sol | max | 75.6% | $0.85 |
| GPT-6 Astra | low | 63% | $0.79 |
| GPT-6 Astra | medium | 67.5% | $1.35 |
| GPT-6 Astra | high | 68.8% | $2.15 |
| GPT-6 Astra | xhigh | 70.5% | $2.45 |
| GPT-6 Astra | max | 70.9% | $4.18 |
| GPT-6.1 Sol | low | 62% | $0.12 |
| GPT-6.1 Sol | medium | 64.5% | $0.22 |
| GPT-6.1 Sol | high | 67.8% | $0.37 |
| GPT-6.1 Sol | xhigh | 69.4% | $0.45 |
| GPT-6.1 Sol | max | 71% | $0.74 |
- Exact score and cost values from the chart data embedded in OpenAI's announcement.