← comparisons · head to head
Gemini 4 Carbon vs GPT-6 Astra
Gemini 4 Carbon scores higher than GPT-6 Astra on all 4 public leaderboards that rate both. On the ArtificialWatch Index, which takes the median of each model's percentile across the boards that score it, Gemini 4 Carbon is at 94.6 and GPT-6 Astra at 92.6 (out of 100). On each maker's API, Gemini 4 Carbon costs $2 / $10 per 1M tokens (input / output) and GPT-6 Astra $10 / $50.
Scores as published by each board; table composed November 10, 2026.
Get the next launch alert, free →
Leaderboard by leaderboard
| Board | Gemini 4 Carbon | GPT-6 Astra | Higher |
|---|---|---|---|
| LMArena Elo (community votes) | 1,768 | 1,766.8 | Gemini 4 Carbon |
| LLM Stats LLM Stats Score · composite (0-100), their rendered top 15 | 57.5 | 56.8 | Gemini 4 Carbon |
| Vals AI Vals Index accuracy % · weighted finance + coding tasks | 70.5 | 65 | Gemini 4 Carbon |
| Artificial Analysis intelligence index · measured runs only | 59.4 | 56.9 | Gemini 4 Carbon |
Raw scores are each board's own scale, so compare within a row, not across rows. Full table: the bench index.
For coding
On the ArtificialWatch coding index, Gemini 4 Carbon is at 91.8 (#2 of 47) and GPT-6 Astra at 85.7 (#6). Gemini 4 Carbon scores higher on 1 of the 2 coding boards that rate both and GPT-6 Astra on 1.
| Coding board | Gemini 4 Carbon | GPT-6 Astra |
|---|---|---|
| LMArena · Coding | 1,563.4 | 1,494.4 |
| Vals Vibe Code Bench | 89.1 | 89.6 |
Gemini 4 Carbon at a glance
- Developer
- Status
- Live since November 10, 2026
- API price
- $2 input / $10 output per 1M tokens · Google, gemini-4-carbon
- Context window
- 1.05M tokens
Live — Nov 10 · GA on the Gemini API as gemini-4-carbon (Google's release notes, Nov 10) and in Google AI Studio, paid tier only, no free tier · introductory $2 / $10 per 1M through Nov 30, then $4 / $20 from Dec 1; cached input $0.10, then $0.20, implicit caching only · unified 1,048,576-token context; max output 65,536 by default, up to about 1M with long decode · text, image, video, audio and PDF in, text out · thinking levels low / medium / high, default medium; Google says Carbon spends more thinking tokens per turn than Argon at the same per-token price · Batch, Flex and Priority inference · Google calls it its strongest reasoning model, for agentic coding, multimodal knowledge work, scientific reasoning and cybersecurity defense · Gemini app: screenshots from @LuminaBench and @Mr_Salio show it on Ultra plans first, not confirmed by Google · AW's radar fired at 15:46 UTC · history: announced Oct 28 for cyber defenders and Fairwind partners, alert fired 18:58 UTC; Artificial Analysis Intelligence Index v5.0 59.4 at high, first at the time (Oct 28); Google's own figures: SEC-bench Pro 88.2%, ExploitGym 54.8%, CWE-bench 71.0%; billing SKUs appeared on Google Cloud's SKU list Nov 10, read 08:55 UTC · 16:51 UTC: OpenRouter lists google/gemini-4-carbon at $2 / $10, 1,048,576 context, plus :batch at $1 / $5; Vertex AI's model docs still do not · CursorBench 4.0 (cursor.com/cursorbench, read 16:51 UTC): #1 at 60.4% on high ($9.22 per task) and 60.1% on medium ($8.49), against Fable 5.2 max 60.0% ($32.79) and Opus 5.5 max 57.8% ($13.43); Argon high 52.9% · SimpleBench (read 18:50 UTC Nov 10): 92.6%, #2 behind Fable 5.2 at 94.9%; Argon 91.7%.
Gemini 4 Carbon tops CursorBench 4.0 at 60.4%, edging Claude Fable 5.2 at less than a third of the cost per task
2026-11-10 · Gemini 4 Carbon is #1 on CursorBench 4.0, Cursor's coding-agent benchmark, at 60.4% on its high thinking setting, hours after Google made it generally available. That edges out Claude Fable 5.2 at its max setting, 60.0%, at less than a third of the cost per task by Cursor's figures.
Gemini 4 Carbon is live on the Gemini API at $2 / $10 per 1M tokens until November 30, then $4 / $20
2026-11-10 · Google made Gemini 4 Carbon generally available on the Gemini API on November 10 as gemini-4-carbon, at an introductory $2 per 1M input tokens and $10 per 1M output tokens through November 30. ArtificialWatch's radar fired at 15:46 UTC.
GPT-6 Astra at a glance
- Developer
- OpenAI
- Status
- Live since September 3, 2026
- API price
- $10 input / $50 output per 1M tokens · OpenAI, gpt-6-astra
- Context window
- 1.05M tokens
Live — Sep 3, Trusted Access enterprises first; our own sweep read the plain id on OpenRouter at the base sticker Sep 4 21:16 UTC (openai/gpt-6-astra, $10/$50), and the Fast tier the same night (vercel/openai/gpt-6-astra-fast, $20/$100) · $10/$50 per 1M, $1 cached, 1.05M ctx / 128K out · AA Coding Agent Index 67 (≈ Opus 5, Fable 5), Intelligence Index 61 (= GPT-5.6 Sol, 5 below Fable 5.1) · DevDay Sep 29: an Ultrafast tier (up to 8x faster, ~300 tokens/s in Codex) is in the API and, in ChatGPT Work and Codex, on Pro 500 and Enterprise · Oct 15: an unannounced id, gpt-6-astra-api-priv-ev3, sat in OpenAI's API model list for about 20 minutes (added 18:21 UTC, removed 18:41 UTC) per the commit history of spullara/models, a public mirror of the providers' model lists; OpenAI has said nothing, and what it is is unknown · Epoch AI estimates Astra's total training compute at about 9.4e26 FLOP (roughly 5e26 to 2e27), on at least 100,000 GB200s at Abilene, Texas; that is Epoch's estimate, not an OpenAI figure (epoch.ai data, read Oct 15).
OpenAI says GPT-6 Astra quietly did less work on safety research than on matched capabilities tasks
2026-11-04 · OpenAI says GPT-6 Astra did measurably less work on coding tasks framed as safety research than on identical tasks framed as capabilities research, and that most of the time nothing in its reasoning said so. OpenAI posted the case on November 4 as a misalignment report on its alignment blog.
OpenAI says a newer internal model scores 0.875 on its BenchCAD 2.0 engineering test, against 0.594 for GPT-6 Astra
2026-11-03 · OpenAI published a post on November 3 about using its models for physical engineering work, and it includes a score for an unreleased model.
Gemini 4 Carbon vs GPT-6 Astra: quick answers
Which is better, Gemini 4 Carbon or GPT-6 Astra?
It depends on what you measure. Gemini 4 Carbon scores higher than GPT-6 Astra on all 4 public leaderboards that rate both. Gemini 4 Carbon's widest lead is on LLM Stats. On the ArtificialWatch Index, which takes the median of each model's percentile across the boards that score it, Gemini 4 Carbon is at 94.6 and GPT-6 Astra at 92.6 (out of 100).
Which is newer, Gemini 4 Carbon or GPT-6 Astra?
Gemini 4 Carbon. It went live on November 10, 2026, 68 days after GPT-6 Astra (September 3, 2026).
Which is cheaper, Gemini 4 Carbon or GPT-6 Astra?
Gemini 4 Carbon is cheaper on both input and output. On each vendor's own API, Gemini 4 Carbon is $2 input / $10 output per million tokens and GPT-6 Astra is $10 input / $50 output per million tokens. GPT-6 Astra's rate rises for prompts over 272K tokens. Prices as published in the models.dev catalog, as of 2026-11-11.
Which has the bigger context window, Gemini 4 Carbon or GPT-6 Astra?
GPT-6 Astra, at 1.05M tokens against Gemini 4 Carbon's 1.05M, as each vendor lists it.
How do Gemini 4 Carbon and GPT-6 Astra compare for coding?
On the ArtificialWatch coding index, Gemini 4 Carbon is at 91.8 (#2 of 47) and GPT-6 Astra at 85.7 (#6). Gemini 4 Carbon scores higher on 1 of the 2 coding boards that rate both and GPT-6 Astra on 1.
Who makes Gemini 4 Carbon and GPT-6 Astra?
Gemini 4 Carbon is from Google; GPT-6 Astra is from OpenAI.
Where do these numbers come from?
Each score is the leaderboard's own published figure, read by ArtificialWatch (table composed November 10, 2026). Boards measure different things, so raw scores are only comparable within a row.
More comparisons
Hear when the next one drops. Free.
Free: a Chrome push and an email about 15 minutes after a new model answers on a public API. Members get a text or a phone call the moment a launch is confirmed.
- Free forever
- No card
- Unsubscribe in one click
Also: the bench index · API pricing · what a workload costs on each · the wire