Claude Sonnet 5 vs Gemini 3.5
Pricing, benchmarks, and use case comparison
Newer versions are out: Claude Sonnet 5 has been replaced by Claude Sonnet 5.5; Gemini 3.5 has been replaced by Gemini 4 Hydrogen.
Verdict
Our pick: Claude Sonnet 5Pick Claude Sonnet 5: it is still served as itself at $2/$10, while Google deprecated Gemini 3.5 Flash on October 8, 2026 and routes its API calls to Gemini 3.6 Flash. For new work, compare Claude Sonnet 5.5 with Gemini 4 Hydrogen.
Specs comparison
| Claude Sonnet 5 | Gemini 3.5 | |
|---|---|---|
| Provider | Anthropic | Google DeepMind |
| Type | Closed source | Closed source |
| Context window | ✓1M | 1,048,576 tokens (Gemini 3.5 Flash; Pro variant not yet released) |
| Input / 1M tokens | $2.00 | ✓$1.50 |
| Output / 1M tokens | $10.00 | $9.00 |
| Release date | 2026-06 | 2026-05 |
Benchmarks
| Benchmark | Claude Sonnet 5 | Gemini 3.5 |
|---|---|---|
| SWE-bench Verified | ~85.2% | - |
| Terminal-Bench 2.1 | 80.4% | - |
| SWE-bench Pro | 63.2% | - |
| OSWorld-Verified | 81.2% | - |
| Terminal-Bench 2.1 (coding) | - | 76.2% |
| MCP Atlas (tool use) | - | 83.6% |
| CharXiv Reasoning (multimodal) | - | 84.2% |
Scores sourced from official provider release posts and independent benchmark aggregators.
Both have successors: Claude Sonnet 5 was followed by Claude Sonnet 5.5 in September 2026, and Gemini 3.5 Flash by Gemini 3.8 Flash and then Gemini 4 Hydrogen, Google's everyday model since November 12, 2026.
Capability and benchmarks
These landed very close. Claude Sonnet 5 posts 80.4% Terminal-Bench 2.1, around 85.2% SWE-bench Verified, and 81.2% OSWorld-Verified for computer use. Gemini 3.5 Flash posted 76.2% Terminal-Bench 2.1, 83.6% MCP Atlas tool use, and 84.2% CharXiv multimodal reasoning. Sonnet 5 has the edge on coding and terminal-agent depth (coding 92 vs 90 on our capability index); Gemini led on multimodal (90 vs 85) and function calling. Beyond Terminal-Bench 2.1 they share no public coding number, so this was a trade-off rather than a clean win.
Status, price, and context
Google deprecated gemini-3.5-flash on October 8, 2026: it left Google AI Studio, and its API requests now route automatically to gemini-3.6-flash. Sonnet 5 is still available on the Claude API, AWS, Google Cloud and Microsoft Foundry, listed by Anthropic as a legacy model with retirement not sooner than June 30, 2027. Pricing has also moved in Sonnet 5's favor: Anthropic made its $2/$10 introductory price permanent, while Gemini 3.5 Flash listed at $1.50/$9. Both have a context of about 1M tokens. Sonnet 5's tokenizer produces about 30% more tokens for the same text than Sonnet 4.6, so its effective cost is higher than the per-token rate suggests.
Which to pick
- Pick Sonnet 5 over Gemini 3.5 Flash for any workload still running on one of them: it is served as itself and has the stronger coding record.
- For new work, compare the successors: Sonnet 5.5 at $2/$10, and Gemini 4 Hydrogen at $1.50/$7.50 per 1M (Gemini 3.8 Flash stays at $0.75/$3.75 through December 31, 2026, then matches that price).
Which should you choose?
Choose Claude Sonnet 5 if...
- →You have a pipeline tuned to Sonnet 5 and want stable behavior while you test Sonnet 5.5
- →You turn thinking off with thinking type disabled or use forced tool_choice, both of which Sonnet 5.5 rejects
- →You need a 1M context window at $2/$10 with the Sonnet 5 request surface
Choose Gemini 3.5 if...
- →You need frontier agent/coding performance without frontier prices
- →Building autonomous agents that make many tool calls
- →High-throughput production workloads that were previously too costly on a Pro model
- →You want a strong default multimodal model with a 1M-token context window