Compare models
Rank every model by cost for the same request, side by side.
Rank every model by cost for the same request, side by side.
| Model | Released | In / Out per 1M | Context | This request | SWE-bench Verified | SWE-bench Pro (Public) | Terminal-Bench 2.0 | Humanity's Last Exam (no tools) | GPQA Diamond | ARC-AGI-2 |
|---|---|---|---|---|---|---|---|---|---|---|
| Claude Haiku 5.5Claude Haiku · Budget | Oct 2026 | $0.10 / $0.50 | 1.0M | $0.015 | — | 64.8% | — | 45.9% | — | — |
| Claude Haiku 4.5Claude Haiku · Balanced | Oct 2025 | $1 / $5 | 200K | $0.150 | 73.3% | — | — | — | 73% | — |
| Claude Sonnet 5.5Claude Sonnet · Balanced | Sep 2026 | $2 / $10 | 1.0M | $0.300 | — | 81.3% | — | 56.9% | — | — |
| Claude Sonnet 5Claude Sonnet · Balanced | Jun 2026 | $2 / $10 | 1.0M | $0.300 | 85.2% | 63.2% | — | 43.2% | — | — |
| Claude Sonnet 4.6Claude Sonnet · Balanced | Feb 2026 | $3 / $15 | 1.0M | $0.450 | 79.6% | — | 59.1% | 33.2% | 89.9% | 58.3% |
| Claude Opus 5.5Claude Opus · Frontier | Sep 2026 | $4 / $20 | 1.0M | $0.600 | — | 89.9% | — | 64.4% | — | — |
| Claude Opus 5Claude Opus · Frontier | Jul 2026 | $5 / $25 | 1.0M | $0.750 | — | 79.2% | — | 56.3% | — | 90.42% |
| Claude Opus 4.8Claude Opus · Frontier | May 2026 | $5 / $25 | 1.0M | $0.750 | 88.6% | 69.2% | — | 49.8% | 93.6% | — |
| Claude Opus 4.7Claude Opus · Frontier | Apr 2026 | $5 / $25 | 1.0M | $0.750 | 87.6% | 64.3% | 69.4% | 46.9% | 94.2% | — |
| Claude Opus 4.6Claude Opus · Frontier | Feb 2026 | $5 / $25 | 1.0M | $0.750 | 80.8% | 53.4% | 65.4% | 40% | 91.3% | 68.8% |
| Claude Opus 4.5Claude Opus · Frontier | Nov 2025 | $5 / $25 | 200K | $0.750 | 80.9% | — | 59.3% | 30.8% | 87% | 37.6% |
| Claude Fable 5.1Claude Fable · Frontier | Sep 2026 | $10 / $50 | 1.0M | $1.50 | — | 81.2% | — | 60.9% | — | 90% |
| Claude Mythos 5.1Claude Mythos · Frontier | Sep 2026 | $10 / $50 | 1.0M | $1.50 | — | 81.2% | — | 60.9% | — | 90% |
| Claude Fable 5Claude Fable · Frontier | Jun 2026 | $10 / $50 | 1.0M | $1.50 | 95% | 80% | — | — | — | 89.2% |
| Claude Mythos 5Claude Mythos · Frontier | Jun 2026 | $10 / $50 | 1.0M | $1.50 | 95.5% | 80.3% | — | 59% | 94.1% | 89.2% |
| Model | Released | In / Out per 1M | Context | This request | SWE-bench Verified | SWE-bench Pro (Public) | Terminal-Bench 2.0 | Humanity's Last Exam (no tools) | GPQA Diamond | ARC-AGI-2 |
|---|---|---|---|---|---|---|---|---|---|---|
| GPT-6 LunaGPT-6 · Budget | Sep 2026 | $0.10 / $0.50 | 1.1M | $0.015 | — | — | — | — | — | — |
| GPT-5.6 LunaGPT-5.x · Budget | Jul 2026 | $0.20 / $1.20 | 1.1M | $0.032 | — | — | — | — | — | — |
| GPT-5.4 nanoGPT-5.x · Budget | Mar 2026 | $0.20 / $1.25 | 400K | $0.033 | — | 52.4% | 46.3% | 24.3% | 82.8% | — |
| GPT-5.4 miniGPT-5.x · Budget | Mar 2026 | $0.75 / $4.50 | 400K | $0.120 | — | 54.4% | 60% | 28.2% | 88% | — |
| GPT-6.1 SolGPT-6 · Balanced | Sep 2026 | $2 / $10 | 1.1M | $0.300 | — | — | — | — | — | — |
| GPT-5.3-CodexGPT-5.x · Balanced | Feb 2026 | $1.75 / $14 | 400K | $0.315 | — | 56.8% | 77.3% | — | 92.6% | — |
| GPT-5.6 TerraGPT-5.x · Balanced | Jul 2026 | $2 / $12 | 1.1M | $0.320 | — | — | — | — | — | — |
| GPT-5.4GPT-5.x · Balanced | Mar 2026 | $2.50 / $15 | 1.1M | $0.400 | — | 57.7% | 75.1% | 39.8% | 92.8% | 73.3% |
| GPT-5.6 SolGPT-5.x · Frontier | Jul 2026 | $4 / $20 | 1.1M | $0.600 | — | — | — | — | 94.6% | 92.5% |
| GPT-5.5GPT-5.x · Frontier | Apr 2026 | $5 / $30 | 1.1M | $0.800 | — | 58.6% | 82.7% | 41.4% | 93.6% | 85% |
| GPT-6 AstraGPT-6 · Frontier | Sep 2026 | $10 / $50 | 1.1M | $1.50 | — | — | — | — | 96% | 95% |
| Model | Released | In / Out per 1M | Context | This request | SWE-bench Verified | SWE-bench Pro (Public) | Terminal-Bench 2.0 | Humanity's Last Exam (no tools) | GPQA Diamond | ARC-AGI-2 |
|---|---|---|---|---|---|---|---|---|---|---|
| Gemini 3 FlashGemini Flash · Budget | Dec 2025 | $0.50 / $3 | 1.0M | $0.080 | 78% | 48.4% | 47.6% | 33.7% | 90.4% | 33.6% |
| Gemini 3.8 FlashGemini Flash · Budget | Sep 2026 | $0.75 / $3.75 | 1.0M | $0.112 | — | — | — | — | — | — |
| Gemini 3.6 FlashGemini Flash · Budget | Jul 2026 | $0.75 / $3.75 | 1.0M | $0.112 | — | 58.7% | — | — | — | — |
| Gemini 2.5 ProGemini Pro · Balanced | Jun 2025 | $1.25 / $10 | 1.0M | $0.225 | 59.6% | — | 32.6% | 21.6% | 86.4% | 4.9% |
| Gemini 3.5 FlashGemini Flash · Balanced | May 2026 | $1.50 / $9 | 1.0M | $0.240 | — | 53.9% | — | — | — | 72.1% |
| Gemini 3.1 ProGemini Pro · Balanced | Feb 2026 | $2 / $12 | 1.0M | $0.320 | 80.6% | 54.2% | 68.5% | 44.4% | 94.3% | 77.1% |
Up and to the left is better value: a higher score for less money on this request.
Benchmarks are the providers' own published numbers (each links to its source), checked 2026-10-10. Settings differ between providers, so compare within a column, not across columns. A dash means no published score.