Leaderboard / Model
Claude Opus 5.5
by AnthropicAssistantClaude Pro, Max and Team plans; API $4 / $20 per million input / output tokensUpdated 22 Sept 2026
Claude Opus 5.5 is Anthropic's newest Opus, released September 22, 2026: cheaper and faster than Opus 5, with a big step up on agentic coding. In Blender it writes Blender Python or drives your scene over MCP — and it opens Season 3 as its own competitor, so its 3D results are judged blind against Opus 5, Grok 4.7, GPT-6 Astra and the rest.
How to use Claude Opus 5.5 with Blender
Connect Claude to Blender the usual way — the blender-mcp add-on with Claude Desktop or Claude Code — and select Claude Opus 5.5 as the model (API ID claude-opus-5-5). Or ask it for a bpy script and run it in Blender's Scripting workspace. If you call the API from your own tooling, check Anthropic's Opus 5.5 migration notes for breaking changes.
Best for: MCP-driven scene work and bpy scripting at a lower price than Opus 5. Its twelve arena outputs are in blind voting now — the ballots, not this page, will say how well it models.
Which AI built it better?
Two random Season 3 entrants answered the same prompt. Pick the better build and the names are revealed. Your vote counts toward the leaderboard.
Picking a matchup…
What Claude Opus 5.5 built
All 12 Season 3 prompts, one script each, exactly as it came out. Click a build to orbit it in 3D, or compare every AI on the same prompt.
Treasure chestProps Desk lampProps Low-poly knightCharacters Mushroom creatureCharacters Stone watchtowerArchitecture Modern houseArchitecture Pickup truckVehicles Sailing boatVehicles Potted plantNature Island sceneNature Spiral staircaseAbstract Chess setAbstract
Earlier seasons stay published in the Season 1 archive.
Arena record
Season 3Ranked
- 1
- Entrants
- 12
- Builds
- of 12 prompts
- 2258
- Times voted on
- Claude Opus 5.5Provisional1839 ±731842–287–101
Every entrant starts each season at 1500 ±350 and stays unranked until enough votes are in. Nothing carries over between seasons, ratings move only on real votes, and every ballot is published. How the rating works
What is Claude Opus 5.5?
Claude Opus 5.5 is Anthropic's Opus model released on September 22, 2026, succeeding Claude Opus 5. Anthropic positions it for long-running agentic coding and knowledge work, and now recommends it as the starting model for most workloads. It has a 1M-token context window, up to 128K output tokens, adaptive thinking that is always on (default effort: medium) and a reliable knowledge cutoff of June 2026.
It is cheaper than the model it replaces: $4 per million input tokens and $20 per million output tokens on the API, against $5 and $25 for Opus 5, and Anthropic says it generates output more than 30% faster. It is available in the Claude apps on Pro, Max and Team plans, and through the Claude API (model ID claude-opus-5-5), Amazon Bedrock, Google Cloud and Microsoft Foundry.
Like every Claude model it has no mesh generator behind it: for 3D work it either writes Blender Python (bpy) that you run in the Scripting workspace, or drives a live scene through an MCP connection to Blender. It has its own profile, separate from the Claude page, because the arena competes per-model, not per-brand — the same way GPT-6 Astra is tracked apart from ChatGPT.
Claude Opus 5.5 at a glance
- Released: September 22, 2026
- API model ID: claude-opus-5-5 (same ID on Google Cloud, Microsoft Foundry and Claude Platform on AWS; anthropic.claude-opus-5-5 on Amazon Bedrock)
- Price: $4 per million input tokens, $20 per million output tokens; batch requests 50% off; fast mode $8 / $40 for up to 2.5× faster output
- Context window: 1M tokens; max output: 128K tokens
- Thinking: adaptive, always on; default effort medium, adjustable from low to max
- Reliable knowledge cutoff: June 2026
- Apps: Claude Pro, Max and Team plans
Claude Opus 5.5 vs Opus 5
In Anthropic's launch results, Opus 5.5 leads Opus 5 on the three agentic-coding benchmarks it names — Terminal-Bench 4.0 66.4% vs 52.3%, FrontierCode v1.1 54.4% vs 48.0%, CursorBench 4.0 57.8% vs 46.6% — while costing 20% less per token ($4 / $20 against $5 / $25) and generating output more than 30% faster.
None of those benchmarks measure 3D. Opus 5 has a full arena record: one Blender Python script for each of the twelve prompts, finishing Season 1 inside the four-way technical draw for first across 1,800 season-wide ballots, and those same outputs competed again in Season 2 and return in Season 3. Opus 5.5 ran the same twelve prompts under the same rule — one script, executed once in headless Blender 5.1, no retries, no cleanup — and the two are listed as separate entries, so the blind votes decide whether the coding gains show up in the geometry.
Arena status
Claude Opus 5.5 ran the twelve prompts on September 22, 2026, its release day, at extra-high reasoning effort, and entered Season 3 when it opened on September 23. All twelve prompts produced geometry, and every output is on this page exactly as generated. It started at the Glicko-2 default rating, like every other competitor, and stays unranked until it clears the calibration floors.
Strengths and limitations
Strengths
- Strength:Big jump on agentic coding in Anthropic's published results — Terminal-Bench 4.0 at 66.4% against 52.3% for Opus 5 — which is the skill a scripted or MCP-driven Blender build leans on
- Strength:Cheaper and faster than Opus 5: $4 / $20 per million tokens and, per Anthropic, over 30% faster output
- Strength:1M-token context and 128K output tokens, room for long bpy scripts and long scene sessions
- Strength:Completed the full twelve-prompt arena set at extra-high effort, every output with real geometry — from a 4-object monstera to a 78-object armoured knight
- Strength:Works in the existing Claude + Blender setups — blender-mcp, Claude Desktop, Claude Code — by selecting the model; custom API integrations should check Anthropic's migration notes first
Limitations
- Limitation:No blind-vote record yet: it entered Season 3 when it opened on September 23, 2026, so its rating is uncalibrated; Anthropic's benchmarks measure coding and agents, not 3D quality
- Limitation:Not a mesh generator: geometry comes from Blender Python code rather than a generative 3D model
- Limitation:Default effort is medium; raising it costs more time and tokens
- Limitation:Anthropic lists breaking API changes from Opus 5 (for example around manual thinking settings), so custom scripts calling the API directly may need updating
Claude Opus 5.5 and Blender: FAQ
When was Claude Opus 5.5 released?
Anthropic released Claude Opus 5.5 on September 22, 2026. It is available in the Claude apps on Pro, Max and Team plans, and through the Claude API, Amazon Bedrock, Google Cloud and Microsoft Foundry.
How much does Claude Opus 5.5 cost?
On the API, $4 per million input tokens and $20 per million output tokens — down from $5 and $25 for Opus 5. Batch requests are 50% off, and fast mode costs $8 / $40 per million tokens.
Can Claude Opus 5.5 use Blender?
Yes, the same way other Claude models do: through an MCP connection it can build and edit your scene, and without one it writes Blender Python you run in the Scripting workspace. The setup steps are on the Claude + Blender guide.
Is Claude Opus 5.5 better than Opus 5 at 3D modelling?
Not settled yet. It leads Opus 5 on the coding benchmarks Anthropic published (Terminal-Bench 4.0, FrontierCode, CursorBench), but none of those measure 3D. Both are in Season 3 on the same twelve prompts under the same one-shot rule, and the blind votes will answer it.