Three-vendor comparison · checked against vendor pages 2026-10-20

Claude Sonnet 5 vs GPT-5.6 vs Gemini 3.1 Pro

As of 2026-10-20: Claude Sonnet 5 is legacy but still Active (retirement not sooner than 2027-06-30, official price $2/$10); OpenAI's flagship is now the GPT-6 family (official price: gpt-6-astra $10/$50, gpt-6.1-sol $2/$10) and GPT-5.6 is no longer the flagship line; Gemini 3.1 Pro lists an official price of $2.00 in and $12.00 out. Anthropic, OpenAI and Google never ran a shared benchmark — so a single 'winner' number does not exist. Here is what is genuinely comparable (price, context, availability) and what is not (leaderboard scores), with the benchmark cells left honestly blank.

#Claude Sonnet 5 #GPT-5.6 #Gemini 3.1 Pro #Honest comparison

The one-line verdict

As of 2026-10-20 there is still no shared benchmark between these three: compare official status, official price and context window, not a single score.

The lineup changed by 2026-10-20: Anthropic marks Claude Sonnet 5 as legacy but it is still Active (retirement not sooner than 2027-06-30, official price $2/$10) and the official docs tell you to consider migrating to Claude Sonnet 5.5; OpenAI's current flagship is the GPT-6 family and GPT-5.6 is no longer in the flagship table; Google's current flagship is Gemini 4 Argon (GA on 2026-10-13) while the Gemini 3.1 Pro compared here is still a preview model. Anthropic published no exact benchmark numbers for Claude Sonnet 5; it only says performance is close to Opus 4.8 qualitatively. OpenAI and Google run their own harnesses on their own dates. So the useful decision axes are official status, official price, context window, output limits, and where each model is actually generally available. Claude Sonnet 5 is a mid-tier, agentic-coding model, and Opus 4.8 still leads the hardest coding, judgment and cyber tasks. Trial all three on your own workload before you decide.

What IS and IS NOT comparable

✅ What you CAN compare

Hard, published facts line up across vendors: list price per million tokens, context-window size, maximum output length, knowledge cutoff, and — crucially — real availability (which model is GA on which platform, and which tiers are still preview). These are the numbers that actually change your bill and your integration plan, and they are the honest basis for a decision.

🚫 What you CANNOT compare

A single 'best model' score does not exist here. Anthropic released no SWE-bench, Terminal-Bench or OSWorld figure for Sonnet 5, and no cross-vendor evaluation was run on identical harnesses, prompts and snapshots. Any table that shows Sonnet 5, GPT-5.6 and Gemini side by side with exact percentages is stitching together third-party estimates from different dates — treat those numbers as directional, never authoritative.

Specs & availability at a glance

Verified specs where vendors published them; benchmark row left blank on purpose.

Attribute	Claude Sonnet 5	GPT-5.6	Gemini 3.1 Pro
Vendor & positioning	Anthropic · mid-tier, agentic coding; now flagged legacy, official docs point to Sonnet 5.5	OpenAI · flagship line is now the GPT-6 family (GPT-5.6 left the flagship table)	Google · flagship is Gemini 4 Argon; the Gemini 3.1 Pro here is still a preview
Context window	1M tokens (default = max)	1,050,000 tokens (same for Sol / Terra / Luna)	Vendor-listed — see Google Vertex docs
Max output	128K (up to 300K via Batches beta)	128,000 tokens (same for Sol / Terra / Luna)	Vendor-listed
Price (input / output per MTok)	$2 / $10 (standard)	Official price: Sol $4/$20 and Terra $2/$12; Sol promotional price until 2026-11-30	Official price: $2.00 input (up to 200k tokens), $4.00 (over 200k); $12.00 and $18.00 output
Knowledge cutoff	January 2026	February 16, 2026	Vendor-listed
Availability / GA status	Active (legacy) — retirement not sooner than 2027-06-30; API, Bedrock, Vertex AI, Foundry, Copilot, OpenRouter	All three GPT-5.6 tiers (Sol / Terra / Luna) have been generally available since July 9, 2026, but the flagship table is now the GPT-6 family (gpt-6-astra, gpt-6.1-sol)	Varies by region & platform
Public benchmark (SWE-bench / Terminal-Bench)	— not vendor-published	— no shared harness	— no shared harness
Best-suited for	Agentic coding at a mid-tier price; Opus-adjacent quality	Trial on your own task mix	Trial on your own task mix

This table was checked on 2026-10-20. Claude Sonnet 5 specs and pricing follow anthropic.com/news/claude-sonnet-5 and the platform.claude.com docs ($2/$10 input/output, confirmed as the standard rate on 2026-08-10; the new tokenizer uses about 30% more tokens on the same text, so it is roughly cost-neutral versus Sonnet 4.6's $3/$15). The GPT-5.6 and Gemini cells now carry each vendor's official price and official specs (OpenAI pricing page, Google pricing page, checked 2026-10-20), not QCode's own pricing. No benchmark number here is Anthropic-published.

Two caveats that break most comparisons online

🔮 GPT-5.6 is no longer the flagship — but still check which mode a benchmark used

As of 2026-10-20 the flagship on OpenAI's official model page is the GPT-6 family (gpt-6-astra is the flagship, gpt-6.1-sol balances intelligence and cost) and GPT-5.6 is no longer in the flagship table; all three GPT-5.6 tiers have been generally available since July 9, 2026, and only older models such as gpt-5.6-sol remain in the official price list. OpenAI's official price list prices Standard, Batch, Flex, Fast and Ultrafast separately, so the same model has a different unit price in each mode, so check the mode before you compare prices; likewise check which mode or compute setting a benchmark measured before you compare scores, because treating one mode's score as the default experience leads to wrong conclusions.

📅 Most 'Sonnet vs Gemini' tables are Sonnet 4.6-era

Claude Sonnet 5 launched on 2026-06-30, so the vast majority of 'Sonnet vs Gemini' comparison tables circulating online were built against Claude Sonnet 4.6 and were never updated. They carry old prices, old context limits and old snapshots. Note that Sonnet 4.6 (claude-sonnet-4-6) is still Active - not retired, tentative retirement no sooner than Feb 17 2027 - while Sonnet 5 is now flagged legacy by Anthropic: the official page says it is still available but that you should consider migrating to Claude Sonnet 5.5, retirement is not sooner than 2027-06-30, and the current Anthropic lineup is Fable 5.1 / Opus 5.5 / Sonnet 5.5 / Haiku 5.5. If a table does not say 'Sonnet 5', assume it is stale; if it says only 'Sonnet 5' with no migration note, it is stale too.

Compare Claude and GPT on one QCode key

Instead of picking a winner from numbers that do not exist, run your own bake-off. A single QCode API key lets you route to Claude Sonnet 5 for agentic coding and to the OpenAI-compatible line for GPT-5.x, so you can A/B the same task and judge on real output, latency and cost.

Claude Code (Claude Sonnet 5)

export ANTHROPIC_BASE_URL="https://api.qcode.cc/api"
export ANTHROPIC_AUTH_TOKEN="$QCODE_KEY"
export ANTHROPIC_MODEL="claude-sonnet-5"
claude

OpenAI Codex CLI (GPT-5.x)

npm install -g @openai/codex
# add QCode profile in ~/.codex/config.toml
codex --profile qcode

Point Claude Code at api.qcode.cc with model claude-sonnet-5, and add a QCode profile to the Codex CLI for GPT-5.x. Then run the identical task through each and compare on your own repo — the only benchmark that actually matters for your decision. QCode does not currently offer the Gemini series; please use Claude, GPT or Chinese models such as claude-sonnet-5-5, gpt-6.1-sol or glm-5.3.

Frequently asked questions

Which of these three is best for coding?

There is no shared, vendor-published benchmark across Anthropic, OpenAI and Google, so no single model can be crowned best for coding from official numbers: Anthropic published no SWE-bench, Terminal-Bench or OSWorld figure for Sonnet 5 and only says qualitatively that its performance is close to Opus 4.8, and Opus 4.8 still leads the hardest coding and judgment tasks. The honest approach is to trial all three on your own repository and evaluate on your task mix, latency and cost rather than a leaderboard. Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days. You can't buy packs while you have an active personal plan; if the plan's quota isn't enough, buy a booster pack.

Is Claude Sonnet 5 better than GPT-5.6?

No one can say from published data. Anthropic did not release exact benchmark numbers for Sonnet 5 and there is no apples-to-apples comparison the two vendors both ran. Sonnet 5 is a mid-tier model priced at $2/MTok input and $10/MTok output (standard rate). Any 'Sonnet 5 beats GPT-5.6' claim you see online is a third-party estimate, not a vendor result.

Can I compare Sonnet 5 and Gemini 3.1 Pro on SWE-bench scores?

Not reliably. Anthropic published no exact SWE-bench, Terminal-Bench or OSWorld number for Claude Sonnet 5 — only the qualitative claim that its performance is 'close to Opus 4.8'. Cross-vendor SWE-bench tables usually mix different harnesses, dates and model snapshots, so they are not directly comparable. Treat any number you see as a third-party estimate, not an Anthropic-published figure. The two models also sit in different places as of 2026-10-20: Anthropic flags Claude Sonnet 5 as legacy while it stays Active (retirement not sooner than 2027-06-30, official price $2/$10), and the official docs point to Claude Sonnet 5.5, while Gemini 3.1 Pro is still a preview model (gemini-3.1-pro-preview) with an official price of $2.00 input (prompts up to 200k tokens) and $4.00 (over 200k), $12.00 and $18.00 output, and Google's flagship is now Gemini 4 Argon (GA on 2026-10-13). Neither vendor publishes aligned SWE-bench, Terminal-Bench or OSWorld numbers, so run each one on the same repository before you decide. QCode does not currently offer the Gemini series.

Is GPT-5.6 generally available?

Yes, still available. As of 2026-10-20 the flagship on OpenAI's official model page is the GPT-6 family (gpt-6-astra official price $10/$50, gpt-6.1-sol official price $2/$10) and GPT-5.6 is no longer in the flagship table: all three tiers have been generally available since July 9, 2026, Sol (gpt-5.6-sol) has an official price of $4/$20 with promotional pricing until 2026-11-30, and Terra (gpt-5.6-terra) has an official price of $2/$12. OpenAI's official deprecation table points gpt-5-mini-2025-08-07 to gpt-5.6-terra and gpt-5-nano-2025-08-07 to GPT-5.6 Luna; those older snapshots are removed from the OpenAI API on 2026-12-11. Claude Sonnet 5 is available via the Claude API, Amazon Bedrock, Google Cloud Vertex AI, Microsoft Foundry, GitHub Copilot and OpenRouter. QCode does not currently offer gpt-6-luna or GPT-5.6 Luna, so please use another GPT model such as gpt-6.1-sol, gpt-6-sol or gpt-5.6-terra.

Run your own bake-off on QCode

One key, real tasks, honest results - decide with your own numbers, not someone else's leaderboard. Not ready for a monthly plan? Pay-as-you-go Packs start at $4 for a one-time quota you use within 7 days. You can't buy packs while you have an active personal plan; if the plan's quota isn't enough, buy a booster pack.

Get a QCode key See pricing