Model comparison
Claude 3 Haiku vs GPT-5.6 Sol
Head-to-head evidence from 12 shared benchmark results across 6 categories. Overall scores shown here use the public BenchAlign v5 ranking lane.
Public leaderboard positions: Claude 3 Haiku #194 (Supported); GPT-5.6 Sol #3 (Supported). Intervals and evidence labels describe ranking uncertainty, not a guarantee for a specific workload.
Evidence parity. Claude 3 Haiku and GPT-5.6 Sol share 12 comparable benchmark results. 0 of 8 categories are comparable. 0 results are unique to Claude 3 Haiku; 34 to GPT-5.6 Sol.
Updated July 23, 2026- Shared results
- 12
- Claude 3 Haiku only
- 0
- GPT-5.6 Sol only
- 34
- Comparable categories
- 0 / 8
Benchmark data for Claude 3 Haiku and GPT-5.6 Sol is coming soon on BenchLM.
Confidence note. This is a partial-evidence comparison with 12 shared benchmark results across 6 evidence categories; 0 of 8 categories currently have scoreable aggregates for both models. Treat the verdict as directional until coverage is more balanced.
Why this result
BenchLM has partial data for these models, but not enough overlapping benchmark coverage to produce a fair score-level comparison yet.
GPT-5.6 Sol is priced at $5.00 input / $30.00 output per 1M tokens, versus $0.25 input / $1.25 output per 1M tokens for Claude 3 Haiku. GPT-5.6 Sol has the larger context window at 1M, compared with 200K for Claude 3 Haiku.
Category breakdown
Exact category averages are shown below. Not measured means BenchLM does not have enough sourced public coverage for that model and category.
| Category | Claude 3 Haiku | Δ | GPT-5.6 Sol |
|---|---|---|---|
| Agentic | Claude 3 HaikuNot measured | MarginNo overlap | GPT-5.6 Sol92.0 |
| Coding | Claude 3 HaikuNot measured | MarginNo overlap | GPT-5.6 Sol64.6 |
| Knowledge | Claude 3 HaikuNot measured | MarginNo overlap | GPT-5.6 Sol94.6 |
| Math | Claude 3 HaikuNot measured | MarginNo overlap | GPT-5.6 Sol87.5 |
| Multimodal | Claude 3 HaikuNot measured | MarginNo overlap | GPT-5.6 Sol83.0 |
Operational comparison
Runtime and commercial metrics are compared only when both models have a complete sourced value.
| Metric | Claude 3 Haiku | GPT-5.6 Sol | Comparison |
|---|---|---|---|
| Input / output priceUSD per 1M tokens | Claude 3 Haiku$0.25 input / $1.25 output | GPT-5.6 Sol$5 input / $30 output | Claude 3 Haiku has the lower combined listed price. |
| Generation speedtokens per second | Claude 3 Haiku138 tok/s | GPT-5.6 SolNot available | A complete speed comparison is not available. |
| First-answer latencyseconds to first token | Claude 3 Haiku1.16 s | GPT-5.6 SolNot available | A complete latency comparison is not available. |
| Context windowmaximum listed tokens | Claude 3 Haiku200K | GPT-5.6 Sol1M | GPT-5.6 Sol lists the larger context window. |
Benchmark Deep Dive
Agentic17 benchmarks
| Benchmark | Claude 3 Haiku | GPT-5.6 Sol | Result |
|---|---|---|---|
| τ²-bench resultsSource | 21.1% | 85.1% | GPT-5.6 Sol leads |
| Terminal-Bench 2.0Source | — | 91.9% | Not comparable |
| BrowseCompSource | — | 92.2% | Not comparable |
| OSWorld 2.0Source | — | 62.6% | Not comparable |
| CyberGymSource | — | 84.5% | Not comparable |
| ExploitGymSource | — | 33.7% | Not comparable |
| ToolathlonSource | — | 58% | Not comparable |
| AA Agentic IndexSource | — | 54.0% | Not comparable |
| GDPval-AASource | — | 61.8% | Not comparable |
| GDPval-AASource | — | 1736 | Not comparable |
| AA BriefcaseSource | — | 1501 | Not comparable |
| AA ITBenchSource | — | 56.2% | Not comparable |
| AA Tau3 BankingSource | — | 33.0% | Not comparable |
| AA AutomationBenchSource | — | 51.2% | Not comparable |
| AA Harvey LABSource | — | 87.2% | Not comparable |
| terminalBenchHardSource | — | 65.9% | Not comparable |
| aaTerminalBench21Source | — | 88% | Not comparable |
Coding8 benchmarks
| Benchmark | Claude 3 Haiku | GPT-5.6 Sol | Result |
|---|---|---|---|
| AA-SciCodeSource | 18.6% | 56.1% | GPT-5.6 Sol leads |
| SWE-bench ProSource | — | 64.6% | Not comparable |
| Terminal-Bench 2.0Source | — | 91.9% | Not comparable |
| deepSweSource | — | 72.7% | Not comparable |
| FrontierCode 1.1 ExtendedSource | — | 60.6% | Not comparable |
| cursorBench32Source | — | 67.2% | Not comparable |
| VulcanBench v3Source | — | 87.0% | Not comparable |
| AA Coding IndexSource | — | 77.4% | Not comparable |
Reasoning4 benchmarks
Knowledge10 benchmarks
| Benchmark | Claude 3 Haiku | GPT-5.6 Sol | Result |
|---|---|---|---|
| Artificial Analysis Intelligence IndexSource | 3.9% | 58.9% | GPT-5.6 Sol leads |
| AA-GPQA DiamondSource | 37.4% | 94.1% | GPT-5.6 Sol leads |
| AA-HLESource | 3.9% | 47.2% | GPT-5.6 Sol leads |
| AA-Omniscience IndexSource | -47.6% | 21.7% | GPT-5.6 Sol leads |
| AA-Omniscience AccuracySource | 17.2% | 58.5% | GPT-5.6 Sol leads |
| AA-Omniscience Hallucination RateSource | 78.2% | 88.8% | Claude 3 Haiku leads |
| GPQASource | — | 94.6% | Not comparable |
| GPQA-DSource | — | 94.6% | Not comparable |
| HealthBench ProfessionalSource | — | 60.5% | Not comparable |
| HealthBench HardSource | — | 33.1% | Not comparable |
Math3 benchmarks
Multimodal3 benchmarks
Inst. Following1 benchmarks
| Benchmark | Claude 3 Haiku | GPT-5.6 Sol | Result |
|---|---|---|---|
| AA-IFBenchSource | 36.1% | 72.7% | GPT-5.6 Sol leads |
Frequently Asked Questions (3)
Can I compare Claude 3 Haiku and GPT-5.6 Sol on BenchLM yet?
Not fully yet. BenchLM is tracking both models, but the sourced benchmark breakdown for this comparison is still coming soon.
Why does this comparison show “coming soon”?
BenchLM only shows category winners and benchmark-level calls when we have sourced results that can be compared fairly. For these models, the public benchmark coverage is not complete enough yet.
What data is available for Claude 3 Haiku and GPT-5.6 Sol today?
Claude 3 Haiku: $0.25 input / $1.25 output per 1M tokens GPT-5.6 Sol: $5.00 input / $30.00 output per 1M tokens Both model pages still include creator, context window, reasoning mode, and other metadata while benchmark coverage fills in.
Related Comparisons
Choose a model with this week’s evidence
Join 2,000+ readers for ranking moves, pricing changes, and the claims that still need proof.
One email each week. Unsubscribe anytime.